Skip to navigation – Site map

HomeIssues35The ArchaeOBIA concept applied to...

The ArchaeOBIA concept applied to data curation for the Acheulean site of Cagny-l’Épinette (Somme Valley, France)

Le concept ArchéOBIA appliqué à la curation des données du site acheuléen de Cagny-l’Épinette (vallée de la Somme, France)
Floriane Peudon and Éric Masson

Abstracts

This paper highlights the promising contribution of the ArchaeOBIA methodology to the toolbox of digital humanities applied to data curation in digital archaeology. ArchaeOBIA stands for the methodological concept of object-oriented image analysis (OBIA) applied to knowledge extraction from archaeological image data. Here, the scope of this concept has been extended for the first time to the data curation of the digitized version of handwritten field drawings. This innovation has been implemented over the accumulated data from the Paleolithic site of Cagny-l’Épinette (Somme Valley, France). After thirty years (1980–2010), the 277 square meters excavated on this site are documented by a heterogeneous corpus of archives, in various analog and digital formats, totaling 23,785 referenced remains. This corpus offered a unique opportunity to take up the methodological challenge of a unique data curation, essential to provide data input for the first exhaustive thematic spatial analysis for the site of Cagny-l’Épinette. This paper discusses the added value and limitations of information reconstruction while building a comprehensive and consolidated prehistorical spatial database, in other words, to build a meta-archaeology of Cagny-l’Épinette. We understand “Meta-archaeology” as a revisit of existing data accumulated over a long period of excavation. Both a validation phase and a data integrity assessment confirm that our curation workflow enables us to optimize the collated data and to reconstruct some of the incomplete data. This is a new inspiring instance of the ArchaeOBIA methodology, which enlarges the scope of applications for the optimization of analog archaeological legacy archives.

Top of page

Full text

This work was supported by the Ministère de l’Éducation Nationale, de l’Enseignement supérieur et de la Recherche, France [3-year doctoral contract, 2016–2019]; the École Doctorale Sciences de l’Homme et de la Société, Université de Lille, France [mobility grants]; and the Laboratoire HALMA ‒ UMR 8164, Université de Lille, France [mobility grants].
We would like to thank the director of the excavation Alain Tuffreau (Professor Emeritus, Université de Lille, France) and his co-director Agnès Lamotte (Professor, Université de Lille, France) for giving us access to the archives of Cagny-l’Épinette. We would like to thank all the archaeologists who worked on and contributed to the production of the data, and in particular Agnès Lamotte, Patrick Auguste (Research Fellow, CNRS, France) and Anne-Marie Moigne (Lecturer, MNHN-CNRS, France) for entrusting us with their databases.
We are grateful to Armelle Masse (Direction de l’Archéologie du Pas-de-Calais, France), scientific officer of the Centre de conservation et d’étude archéologiques du Pas-de-Calais and to Mickaël Courtiller (DRAC Hauts-de-France, France), documentary researcher at the Centre de documentation patrimoniale, for respectively providing us with access to the collections and excavation reports of Cagny-l’Épinette.
We would like to express our gratitude to the Laboratoire TVES Lille ‒ ULR 4477 (Université de Lille, France) for access to Adobe Illustrator and eCognition software licenses.
The thesis work from which this article has been prepared received the Prix Scientifique L’Harmattan in 2023. Our most sincere thanks to them.
We also thank the reviewers whose constructive comments helped improve our manuscript.

1 | Introduction

1The destructive nature of archaeological excavations has led archaeologists to an ever-increasing desire to acquire more data in the field, a frenzy accentuated by the process of computerization (McVicar 1986). This statement remains valid today, as evidenced by the accumulation of data that accompanies the evolution of scientific practices and digital tools, without necessarily replacing traditional methods. On active excavations over several decades, this progressive computerization has led to a proliferation of recordings in a wide variety of analog and digital formats. This heterogeneity therefore becomes a major scientific challenge for the production of knowledge syntheses from large datasets.

2In this context, designing, structuring and populating a digital database is a complex task, since it raises the question of its architecture, the relevance of the information to be integrated, its degree of integrity, its interoperability, its durability and its use by different users. The modelling of an archaeological database has therefore called for a number of methodological considerations since the earliest designs of data and dissemination of archaeological knowledge infrastructures (e.g., Cheetham and Haigh 1992; Chenhall 1971, 1968, 1967; Cowgill 1967; Dibble 1987; Dibble and McPherron 1988; Gardin 1974, 1971; Gardin and Garelli 1961; Ginouvès 1971; Ginouvès and Guimier-Sorbets 1978; Hodder 1997; Ryan 1992), and more recently in the context of open data (e.g., Bird et al. 2022; Esteva et al. 2010; Fronza 2016; Fronza et al. 2003; Kintigh 2006; Le Goff et al. 2015; Marlet et al. 2022; McManamon and Kintigh 2010; Rabinowitz et al. 2016; Ravindranathan et al. 2004; Richards 2004, 1997; Ryan 2004). Georeferenced data has played a key role in these considerations, for the creation of inventory cartographies (geovisualization) and the exploitation of the spatial and attribute properties of archived objects (spatial analysis) (e.g., Huggett and Ross 2004; McKeague et al. 2019, 2012). Thus, the development of custom-built software has accompanied the increasing use of Geographic Information System (GIS) solutions (e.g., Barge et al. 2008; Bernard 2019; Buchsenschutz et al. 1986; Buchsenschutz and Debanne 1978; Dibble and McPherron 1988; Fronza et al. 2003; Gorton et al. 2006; Husi and Rodier 2011; Kandel et al. 2023; McKeague and Jones 2008; Murray 2004; Niccolucci and Richards 2019; Pirot et al. 2008; Tennant 2007; Willmes et al. 2017).

3With thirty years of heterogeneous archives (1980–2010), 23,785 referenced remains and 277 square meters excavated, the site of Cagny-l’Épinette (Somme Valley, France) offered an excellent opportunity for methodological research on digital tools and the assessment of their scope of application in optimizing the use of analog legacy archives. Its archives are the result of evolving field recording protocols and of numerous studies carried out by successive teams at the site (Tuffreau et al. 1986, 1995; Dibble et al. 1997; Tuffreau and Marcy 2002). Thus, this case study falls within Kulasekaran et al. (2014) statement: “For research projects midway between the “long tail” [referencing to the “long tail of science and technology” defined by Wallis et al. (2013)] and the new data model, the challenge is to merge old and new practices, to shape legacy data into new systems without losing meaning and without overwriting the processes through which data were conceived.” Due to its time-consuming nature, the collation and processing of all the documents from Cagny-l’Épinette had never been carried out before. Our paper therefore proposes to take up the methodological challenge of curating data from thirty years of archives. The aim is to obtain a consolidated (i.e., filling data gaps), enriched (i.e., new data attributes), interoperable dataset (i.e., shapefile format) for the first exhaustive thematic spatial analysis of the site of Cagny-l’Épinette (Peudon 2021; Peudon et al. 2021). We believe that revisiting archived data with a different methodological perspective belongs to the concept of meta-archaeology.

4Due to the nature of the documents and the protocol adopted, this challenge lies at the crossroads of four distinct approaches: (a) redocumentarization, (b) data reuse, (c) new data acquisition and (d) data integrity control.

5Regarding analog archives, the preliminary step of document digitization (i.e., scanning, vectorization, manual editing) falls within Salaün’s (2007) definition of redocumentarization, i.e., the transposition of traditional documents onto a digital medium with the creation of metadata.

6Data re-use refers to the use of data collected by third parties. In our project, this means collating both primary and processed archives to carry out an exhaustive site-wide analysis.

7The acquisition of new data stems from our protocol mobilizing a CAD ➔ GIS ➔ ArchaeoOBIA ➔ GIS-type processing workflow (Peudon 2021; Peudon et al. 2021). Here, CAD refers to Computer-Aided Drawing. ArchaeOBIA refers to the object-oriented image processing and analysis method (OBIA) proposed by Lamotte and Masson (2016) and Masson and Lamotte (2018). Illustrator (Computer-Aided Drawing), ArcGIS (GIS) and eCognition (ArchaeOBIA) have contributed to this processing workflow. Here, the proposed methodological innovation consists in using eCognition for the first time on a corpus of field drawings sourced from handwritten documents. Until now, this software has mainly been reserved for remote sensing imagery (e.g., Davis et al. 2024; De Laet et al. 2007; Freeland et al. 2016; Jahjah et al. 2007; Mohlehli et al. 2023; Trier et al. 2009; Verhagen and Drăguţ 2012).

8According to Atici et al. (2013), “Data integrity refers to the internal consistency and structural coherence of a dataset, as well as data quality issues”. Following this principle, one of the aims of our project (Peudon 2021) was to perform a data integrity check in response to the need to assess data quality prior to any digital analyses (Bennett et al. 1984; Wong and Lee 2005), a concern that remains relevant today (e.g., Bird et al. 2022). This control was all the more necessary as, in the case of Cagny-l’Épinette, the final information is an aggregation of multi-source data.

2 | Presentation of the site and background to our research project

9The Acheulean open-air site of Cagny-l’Épinette is located at the confluence of the Somme and Avre rivers (fig. 1, A). It belongs to the alluvial formation IV of the stepped terrace system of the Middle Somme Valley (Antoine 1990; Haesaerts et al. 1984; Haesaerts and Dupuis 1986; Tuffreau et al. 1986, 1982). Its stratigraphic sequence is composed of a fluvial sequence ‒ whose coarse- and fine-grained fluvial deposits are correlated to MIS (Marine Isotope Stages) 10 and 9 respectively ‒ overlaid by a silty loess cover whose first sandy silt deposits are dated to MIS 8 (Antoine 1990; Antoine and Tuffreau 1993; Bahain et al. 2001; Bates 1993; Bourdier et al. 1974; Haesaerts et al. 1984; Haesaerts and Dupuis 1986; Laurent 1993; Laurent et al. 1998, 1994; Tuffreau 1989; Tuffreau et al. 1995, 1986, 1982; Van Vliet-Lanoë 1989).

Figure 1. |A| Location of the site of Cagny-l’Épinette (black star annotated “Ep”) in the Somme basin. |B| Density map of the inventoried remains (levels H to J), expressed as a number of remains per square meter (Map of France: Modified from IGN 2016 Open source; DEM data: RGE ALTI® 5 m (© IGN), modified; Hydrographic network map: BD CARTHAGE® (© IGN), modified; Density Map: F.l Peudon).
|A| Localisation du site de Cagny-l’Épinette (étoile noire annotée « Ep ») dans le bassin de la Somme. |B| Carte de densité des vestiges inventoriés (niveaux H à J), exprimée en nombre de vestiges par mètre carré (Carte de France : IGN 2016 Open source, modifiée ; Données MNT : RGE ALTI® 5 m (IGN), modifié ; Carte du réseau hydrographique : BD CARTHAGE® (© IGN), modifiée).

Figure 1. |A| Location of the site of Cagny-l’Épinette (black star annotated “Ep”) in the Somme basin. |B| Density map of the inventoried remains (levels H to J), expressed as a number of remains per square meter (Map of France: Modified from IGN 2016 Open source; DEM data: RGE ALTI® 5 m (© IGN), modified; Hydrographic network map: BD CARTHAGE® (© IGN), modified; Density Map: F.l Peudon). |A| Localisation du site de Cagny-l’Épinette (étoile noire annotée « Ep ») dans le bassin de la Somme. |B| Carte de densité des vestiges inventoriés (niveaux H à J), exprimée en nombre de vestiges par mètre carré (Carte de France : IGN 2016 Open source, modifiée ; Données MNT : RGE ALTI® 5 m (IGN), modifié ; Carte du réseau hydrographique : BD CARTHAGE® (© IGN), modifiée).

10Following the discovery of archaeological remains during quarry operations in Cagny (Agache 1971; Bourdier et al. 1974), a 30-year-long excavation led to the discovery of thousands of flint artifacts and faunal remains within the alluvial sequence (Levels I, I0, I1, I1A/IB, I1B, I2, I/J, J) and the first deposits of the silty loess cover (Levels H, H1, Hx) (fig. 1, B) (Tuffreau et al. 1995, 1986, 1982). The essential aspect of the site lies in the well preserved lithic and faunal assemblages embedded in the alluvium, and more particularly in the faunal remains showing evidence of carcass processing by hominins (Auguste et al. 2005; Moigne 1989, 1988; Moigne et al. 2016; Peudon 2021; Tuffreau et al. 1995, 1986). Cagny-l’Épinette thus yields a key dataset in the process of building our knowledge regarding both hominin activities at a site scale and Acheulean regional settlement in northern France during the MIS 9 (Tuffreau 2005; Tuffreau et al. 2008, 1997, 1986).

11Hence, computer intra-site spatial analyses have been performed to apprehend some specific taphonomic and palethnological aspects of the faunal and lithic assemblages embedded within the alluvium (Deusy 2003; Dibble et al. 1997; Drubay 2003; Lefevre 2003; Léopold 1993; Matton 2007; Robiolle-Gautier 1992; Spellemaeker 2003; Tuffreau et al. 1995, 1986). Yet, as the excavation progressed, each of them regarded a restricted part of the overall available final data set.

12In order to exploit the data to its full potential, a new approach was considered at the entire site scale (Peudon 2021, 2013; Peudon et al. 2021). The first research line of this project aimed to collate these complex archives into a coherent and consolidated dataset. Given the characteristics of its background data, the site of Cagny-l’Épinette was chosen for an innovative methodological research project regarding the optimization of an integral archives management and data processing.

13Thus, this paper aims to highlight the potential of pairing a Geographic Information System software (ArcGIS) with an Object Based Image Analysis software (eCognition®). This combination of tools represents a robust processing workflow to link spatial information with qualitative data, both extracted from diversified and scattered data generated by decades of field and individual laboratory works. This new protocol was thoroughly extended to the overall 277 square meters excavation (Peudon 2021; Peudon et al. 2021).

3 | From GEOBIA to ArchaeOBIA

14The object-oriented image analysis is a scientific and technical field of digital image processing that originates from the development of tools for automatic shape recognition in the late 1960s (Prewitt 1970; Fu and Mui 1981) and for digital image segmentation in the late 1970s (Fu and Mui 1981; Pal and Pal 1993; Zhang 1996). Since the mid–1990s, the need to process image data directly at the scale of the objects of interest, the image objects (e.g., Hay et al. 1997, 1996), has emerged as a necessary alternative to pixel-by-pixel digital processing approaches.

15In 2006, the name OBIA (Object Based Image Analysis) was proposed by Hay and Castilla (2006), followed by GEOBIA (GEographical Object Based Image Analysis) (Hay and Castilla 2008). The object-based approach, which emerged in Earth Observation within Geosciences in the 1990s, really took off in the early 2000s, with the commercial availability of high-resolution remote sensing data (Baatz et al. 2008; Baatz and Schäpe 2000; Blaschke 2010, 2003; Blaschke and Lang 2006; Blaschke and Strobl 2001; Hay et al., 2005, 2001; Lang and Blaschke 2006; Willhauck 2000). At the same time, the introduction of software for computer-aided segmentation and classification, e.g., eCognition in 2000, has contributed significantly to this impetus. This software is one of the main methodological tools that operationally support the GEOBIA paradigm at international level. Indeed, eCognition processes image data as a set of discrete objects, described by their geometric, morphological, topological and semantic attributes (Baatz et al. 2008; Baatz and Schäpe 2000; Blaschke 2003; Blaschke and Strobl 2001; Hay et al. 2005, 2001; Willhauck 2000).

16The application of the object-oriented method, part of the broader domain of Machine Intelligence (Davis 2020), appeared later in archaeology (Jahjah et al. 2007; De Laet et al. 2007; Davis 2018). It is recognized as a key tool in predictive analysis for creating maps of archaeological potential (Verhagen and Drăguţ 2012) and in digital prospection (Freeland et al. 2016) for providing the necessary data for the survey and preventive protection of cultural heritage, and for archaeological mapping (e.g., Cerrillo-Cuenca 2017; Davis et al. 2024, 2019; De Guio et al. 2015; Inomata et al. 2017; Magnini et al. 2024; Mohlehli et al. 2023; Pregesbauer et al. 2014; Sărășan et al. 2020; Sevara et al. 2016; Trier et al. 2009; Witharana et al. 2018).

17Although object-oriented methodology has been used primarily in landscape archaeology (Davis 2018), it represents an evolution of the pixel-by-pixel approach already used in the past at the scale of the archaeological object (e.g., Goodson 1989; Grace et al. 1985; McPherron 1991). This involved extracting object-oriented information, i.e., geometric and shape measurements, from film photographs (Grace et al. 1985), video images and scans of archaeological drawings (Goodson 1989), video images (McPherron 1991) or 3D models (Loriot et al. 2007). The GEOBIA method comes therefore as a new tool in a long-standing problem of optimizing information extraction. This new paradigm has been applied to microphotographs of thin sections (Hofmann et al. 2013) or digital photographs (Lamotte and Masson 2016; Masson and Lamotte 2018). The results obtained led Lamotte and Masson (2016) to propose the term ArchéOBIA (ArchaeOBIA) to designate the application of the GEOBIA method in the domain of ArchaeoSciences. Extending the scope of the concept to “the application of object-based image analysis to archaeological research, irrespectively of the scale of investigation”, Magnini and Bettineschi (2019) propose a theoretical reference workflow for ArchaeOBIA projects, aimed at improving the interoperability of processing protocols and the comparability of results.

18Following on from these works, Peudon (2021) proposes a novel application of the ArchaeOBIA method to CAD data derived from handwritten field drawings of archaeological remains. Prior to this research, no (Archae)OBIA processing had yet been used on this type of scientific material. For spatial analysis purposes, the aim was to produce a curation and enrichment of data from the handwritten archives. Complementing this objective, the implementation of the ArchaeOBIA processing enabled us to combine the subjectivity of the expertise of the archaeologist with the objectivity of computerized measurements.

19Our approach therefore differs from other archive digitization projects at other Paleolithic sites, which also have heterogeneous, analog and digital documentation. In these cases, in response to the necessity of reconciling analog and digital archives, the data curation approach adopted is often limited to the design of a database (semantic data) and a GIS (geometric data). An information system was designed at the Solutrean site in Fressignes (France) (Bouyssi et al. 2012; Houllier and Arnoux 2001). In this case, a relational database was populated with both data from excavation notebooks and recent data. Field drawings were vectorized using GIS software. Moreover, a specific acquisition tool, i.e. FrAcTool (Fressignes Acquisition Tool), was developed to populate the database in the field. Other examples of projects involve the design of a database compiling descriptive semantic information (from handwritten documents and/or digital files), and a GIS compiling spatial information (point features and/or vectorized handwritten field drawings), both being joined based on a common attribute. Such work was carried out at the Italian site of Isernia La Pineta (Middle Pleistocene) (Gallotti 2004) and at the site of Sainte-Anne I (Middle Paleolithic, France) (Santagata et al. 2007). Some projects were carried out as part of larger international collaborations. The work of D’Andrea et al. (2002) and Gallotti & Piperno (2003), as part of the Italian Archaeological Mission, on the Ethiopian Oldowan sites of Garba IV, Gomboré I, and Karre are among these projects. Another example of a collaborative project is the design of the GIS for the Acheulean sites of Thomas I Quarry and Oulad Hamida 1 Quarry (Morocco) as part of the Franco-Moroccan “Casablanca” project (Gallotti et al. 2011).

4 | Innovating in Digital Archaeology of Paleolithic

4.1 | Archives diversity and quality

20The archives of the site of Cagny-l’Épinette consist of field and post-excavation documentations, both including analog and digital formats. Abundant diversified archives have thus been compiled through an upstream gathering work (Peudon 2021; Peudon et al. 2021). They comprise several categories of documents: field drawings, field recordings of the measurements of the coordinates, field photographs along with inventories and digital analytical databases of the archaeological remains (see Peudon (2021) for a detailed presentation of the archives). These categories show themselves variations of analog documents and digital files. On the one hand, analog documents encompass notebooks and individual lists; some are handwritten, while others are printed, computerized or half-computerized. On the other hand, various software programs were used to generate the digital files, thus including several formats to deal with. Furthermore, data model varies in structure (both number and type of attributes) within each document category.

21The field recording protocol of the spatial information included mapping and measurement of the coordinates (x, y, z) of the excavated remains (Dibble et al. 1997; Tuffreau et al. 1995, 1986; Tuffreau and Marcy 2002). The archives are comprised of hundreds of one tenth scale planimetric drawings of the remains. This field drawing protocol was sustained per square meter until 2002 and was then discontinued from 2003 to 2010. Part of these drawings was manually composed on graph paper. While others were part of a semi-manual process, by highlighting the remains on vertical photographs, either directly on them, or on tracing paper. All field drawings display control points, enabling accurate georeferencing. From 1980 to 1990, the manual coordinate measurements (x, y, z) of the archaeological remains were registered in annual handwritten field books. The central position of each artifact was described by one set of (x, y, z) coordinates. From 1991 to 2010, the data acquisition involved a total station, whose original computerized recordings were yearly archived in Word files or Excel files. The process of measurement was changed for the remains with a longitudinal axis, which were then described by three sets of (x, y, z) coordinates (i.e., tip and central positions).

22The comparison of the recordings from both previous protocols with the other archives highlighted missing coordinates for some remains (Peudon 2021; Peudon et al. 2021). Thus, gathered data ranges from complete sets (x, y, z) to sets bereft of coordinates (-, -, -), including incomplete sets (i.e., (x, y, -) and (-, -, z)). Consequently, the spatial information available is uneven.

23The descriptive data of the remains result from lithic typo-technological (Lamotte 2012, 2001, 1999; Lamotte and Fabre 2007; Lamotte and Tuffreau 2016, 2001; Léopold 1997, 1993, 1989; Tuffreau et al. 1995, 1986; Tuffreau and Ameloot-Van der Heijden 1991) and archaeozoological (Auguste et al. 2005; Moigne 1989, 1988; Moigne et al. 2016; Tuffreau et al., 1995, 1986) expertise. Most of these data have been typewritten in numerous distinct Excel spreadsheets. Moreover, handwritten books encompass a part of the typo-technological information from the artifacts study. Depending on the specialist, lithic artifacts were characterized by up to some 20 attributes including measurements, typo-technological features, raw materials, refittings, mechanical and chemical weathering processes or use-wear observations. The faunal remains were described by up to 25 attributes including measurements, anatomical and taxonomic determinations, age class, refittings, natural surface alterations or anthropogenic marks. Regarding the morphometric information, length (L), width (W) and thickness (THK) were measured with a caliper. Lithic artifacts smaller than 20 mm take the generic length value “20” (with the annotation “< 20 mm”), with no indication of the exact measurement, in the database. As for the very large faunal remains, for which calipers are unsuitable, they were roughly estimated. Moreover, the recording strategy varied according to the type of remains and over time. This results in different degrees of data availability, with complete (L, W, THK), incomplete (L, W, -) and (L, -, -) sets, or no data at all (-, -, -).

24Therefore, decades of excavation campaigns and post-excavation studies have yielded complex data — extensive, complementary, sometimes redundant and, to a certain extent, incomplete — obviously calling for multisource data curation (Peudon 2021; Peudon et al. 2021).

4.2 | From data curation to knowledge extraction: a GIS/ArchaeOBIA workflow

25The scientific challenge was to design and augment a spatial database that comprises for the archaeological remains (1) semantic data (descriptive attributes) and (2) geometric data (point and polygon features) (Peudon 2021; Peudon et al. 2021).

26The different formats and contents of the input archives necessitated the development of a tree-based protocol involving several processing workflows run simultaneously (fig. 2). The processing of field drawings was the most complex (Peudon 2021, 2013; Peudon et al. 2021). It comprises several steps: (1) scanning, (2) vectorization, (3) georeferencing, (4) object-oriented analysis, (5) mosaicking of shape files for each archaeological level.

Figure 2. Cagny-l’Épinette — Processing protocol applied to the field drawings (left) and the descriptive data (right) of the archaeological remains, with the software used for each step. See description in text (Modified from Peudon (2021)).
Cagny-l’Épinette Protocole de traitement appliqué aux relevés de terrain (à gauche) et aux données descriptives (à droite) des vestiges archéologiques, avec les logiciels sollicités à chaque étape. Voir description dans le texte (Modifié d’après Peudon (2021)).

Figure 2. Cagny-l’Épinette — Processing protocol applied to the field drawings (left) and the descriptive data (right) of the archaeological remains, with the software used for each step. See description in text (Modified from Peudon (2021)). Cagny-l’Épinette — Protocole de traitement appliqué aux relevés de terrain (à gauche) et aux données descriptives (à droite) des vestiges archéologiques, avec les logiciels sollicités à chaque étape. Voir description dans le texte (Modifié d’après Peudon (2021)).

27Two alternatives have been developed (see 2A–3A and 2B–3B in fig. 2). They differ in the choice of software (CAD software Illustrator CS3 or GIS software ArcMap 10.7) used for vectorization and, consequently, in the order of the vectorization and georeferencing steps. All files output from steps 2A–3A and 2B–3B have been exported in GeoTIFF format, which contains georeferencing information, for further processing; these are RGB-coded files with only 3 colors: red (fauna), black (lithic artifacts and flint nodules) and white (image background). The next step involved GEOBIA (eCognition Developer, Trimble), for its ability to perform automatic and accurate digitization, as the main purpose was to develop a method to process a large set of already existing rasterized vector files into GIS-interoperable data. Interoperability between CAD and GIS files remains indeed an ongoing technical issue for any archaeologist working with datasets encompassing a significant amount of already existing CAD archives. This is the first advantage of our methodology, as for a long time, CAD, in various formats, was used to vectorize field drawings, without however enabling spatial analysis, unlike our processing workflow, which enables further GIS processing.

28The second advantage of GEOBIA processing lies in the automatic calculation of attributes. Applying the same attribute calculation algorithms to all GeoTIFF files, even to GIS vector data from step 2A–3A that did not require automatic digitization by eCognition, ensured consistency and consolidation of the final database. GEOBIA thus enabled us to solve the problem of some incomplete spatial information, i.e., missing geographic coordinates. GEOBIA has also made it possible to enrich morphometric attribute data, still uncommon in archaeology for the description of remains (e.g., surface, perimeter, shape indices, compactness index). In detail, this final step was carried out using the Developer version of eCognition. It is based on an object hierarchy derived from image data segmentation, which is integrated into a rule set comprising an algorithmic processing sequence (Baatz and Schäpe 2000; Trimble 2016; Willhauck 2000). An image is therefore reorganized into image objects, all of which can be described with attributes derived from the pixels’ RGB values and their spatial distribution (dimensions, geometry, texture, etc. of the image objects). Once exported in shapefile format, these image objects can then be easily used in a GIS interface (Baatz and Schäpe 2000; Blaschke 2010, 2003; Blaschke and Strobl 2001; Hay et al. 2005; Hay and Castilla 2006; Trimble 2016; Willhauck 2000).

29We have therefore developed a classification rule set that automatically segments the digitized versions of the drawings of Cagny-l’Épinette to extract image objects classified in the faunal and lithic categories (fig. 3; suppl. S1) (Peudon 2021; Peudon et al. 2021).

30The classification rule set developed in this work is as follows, with the specifics of each algorithm taken from Trimble (2016) (fig. 3; suppl. S1):

  1. Input data (RGB-coded rasterized vector data in GeoTIFF format) (fig. 3, A–C).

  2. Initial segmentation of the raster image using the Quadtree Based Segmentation algorithm (fig. 3, D). Based on RGB input data, this top-down segmentation method divides the image into homogeneous quadrants. As remains perimeter geometry becomes more complex, this algorithm goes down to pixel level to delineate the remains border with high segmentation accuracy. Segmentation is complete when all quadrants, regardless of their dimensions, comply with the homogeneity criterion (i.e., scale parameter), i.e., when the maximal color difference within each quadrant is lower than the threshold. In our study, the homogeneity criterion was set to the lowest possible value 1 for pixel-accurate delineation, thus ensuring optimal object extraction with respect to the native quality of the source CAD file. The quadtree segmentation technique (Finkel and Bentley 1974; Hunter and Steiglitz 1979; Samet 1979) has proven to be efficient and relevant for multiscale image processing, from grayscale images (e.g., Chang et al. 1997; Hunter and Steiglitz 1979; Shusterman and Feder 1994) to high-resolution imagery (e.g., Perrolas et al. 2022; Sun et al. 2024). This robust method performs well in a short processing time, particularly in the case of contrasting image objects, which is the aim of our field drawing digital processing.

  3. Object classification by membership function using the basic Classification algorithm (fig. 3, E). Each segment is assigned to one of the “Lithic”, “Fauna” or “Image background” classes. The features of each class were defined beforehand using the values of the RGB layers of the image, i.e. red for faunal remains (R = 227, G = 5, B = 19), black for lithics and flint nodules (R = 0, G = 0, B = 0) and white for background (R = 255, G = 255, B = 255). Furthermore, as each chosen color is defined by a different red (R) value, its respective RGB criterion was based solely on the value of the latter.

  4. Object merge by class using the Merge Region reshaping algorithm (fig. 3, E–F). This classification-based segmentation algorithm merges neighboring quadrants of the same class into larger image objects. Each image object now matches a bone or lithic artifact or a flint nodule.

  5. Export of the output data in shapefile format with attributes using the Export Vector Layer algorithm (fig. 3, G). In this case, geometry (e.g., length, width, orientation) and position (i.e., xy coordinates) object features and one class-related feature (i.e., relations to classification feature) were selected (suppl. S2). Furthermore, data were calculated and exported in 2 dimensions, as the images used are 2-dimensional; Elevation data (z) from Cagny-l’Épinette were not sufficient for such calculations in real 3D.

Figure 3. Cagny-l’Épinette — ArchaeOBIA processing of the digitized versions of the field drawings; example of the area [20–21/Q-S]. |A| Photographs, |B| Field drawing; |C| Rasterized vector data (GeoTIFF format), superimposed remains in dashed line are processed on a separated vector layer. |D to G| — Steps of the classification rule set applied in eCognition: |D| Initial segmentation, |E| Classification of the “Fauna”, “Lithic” and “Background” objects and merging of the “Fauna” and “Lithic” objects, |F| Merge of the “Background” object, |G| Export in shapefile format with an attribute table (|A| A. Tuffreau, |B| G. Leroy, |C to G| Fl. Peudon; Modified from Peudon (2021)).
Cagny-l’Épinette — Traitement ArchéOBIA des vectorisations des relevés de terrain ; exemple du secteur [20-21/Q-S]. |A| Photographies ; |B| Relevé de terrain ; |C| Données vectorielles rastérisées (format GéoTIFF), les vestiges superposés en tireté sont traités sur une autre couche vectorielle ; |D à G| Étapes de la règle de traitement appliquée dans le logiciel eCognition : |D| Segmentation initiale, |E| Classification des objets « Faune », « Lithique » et « Fond » et refaçonnement des objets « Faune » et « Lithique », |F| Refaçonnement de l’objet « Fond », |G| Export au format de fichier de formes avec sa table attributaire(|A| A. Tuffreau, |B| G. Leroy, |C à G| F. Peudon ; Modifié d’après Peudon (2021)).

Figure 3. Cagny-l’Épinette — ArchaeOBIA processing of the digitized versions of the field drawings; example of the area [20–21/Q-S]. |A| Photographs, |B| Field drawing; |C| Rasterized vector data (GeoTIFF format), superimposed remains in dashed line are processed on a separated vector layer. |D to G| — Steps of the classification rule set applied in eCognition: |D| Initial segmentation, |E| Classification of the “Fauna”, “Lithic” and “Background” objects and merging of the “Fauna” and “Lithic” objects, |F| Merge of the “Background” object, |G| Export in shapefile format with an attribute table (|A| A. Tuffreau, |B| G. Leroy, |C to G| Fl. Peudon; Modified from Peudon (2021)). Cagny-l’Épinette — Traitement ArchéOBIA des vectorisations des relevés de terrain ; exemple du secteur [20-21/Q-S]. |A| Photographies ; |B| Relevé de terrain ; |C| Données vectorielles rastérisées (format GéoTIFF), les vestiges superposés en tireté sont traités sur une autre couche vectorielle ; |D à G| Étapes de la règle de traitement appliquée dans le logiciel eCognition : |D| Segmentation initiale, |E| Classification des objets « Faune », « Lithique » et « Fond » et refaçonnement des objets « Faune » et « Lithique », |F| Refaçonnement de l’objet « Fond », |G| Export au format de fichier de formes avec sa table attributaire(|A| A. Tuffreau, |B| G. Leroy, |C à G| F. Peudon ; Modifié d’après Peudon (2021)).

31On completion of the GEOBIA processing, 932 output shape files were related with an attribute table detailing the geometric and spatial properties of the remains. Using spatial and/or attribute joins in ArcGIS, these data were then supplemented with the archaeological descriptive data.

32As a result of this data curation, two GIS vector formats are now available at the scale of the excavation: polygons for remains with a field drawing, and points for all remains with geographic coordinates (field measurements and GEOBIA object centroids). The final version of the curated database is thus available in the interoperable shapefile format.

5 | Results and discussion: expertise of a curated prehistorical database

5.1 | Results from our ArchaeOBIA-GIS methodology

33For the 932 digitized versions of field drawings processed by ArchaeOBIA, we obtained the segmentation of 13,077 archaeological remains (i.e., 54.9% of all inventoried remains) with geographic coordinates and morphometric attributes (Peudon 2021). As for the remaining 45.1%, they are either outside the area of application of the field drawing protocol, or their drawing is missing.

34Among the remains processed by ArchaeOBIA, 1,788 bone remains, lithic artifacts and flint nodules (i.e., 7.5% of the remains), spread over 147 square meters, had their missing xy Cartesian coordinates reconstructed (fig. 4, suppl. S2). In 19 of these square meters, the consolidation reaches + 50% to + 100% of remains that are now valid and reliable for spatial analysis in the XY plane (fig. 4, A). For the remaining 128 square meters, consolidation rates range from + 0.4% to + 47.8%. In rare cases, where only the elevation coordinate (-, -, z) was entered, the ArchaeOBIA method allowed us to reconstruct the xy geolocation (fig. 4, B). In the 10 square meters concerned by these consolidations (+ 1.4% to + 17.6%), we have achieved high final rates of remains valid and reliable for spatial analysis in the XYZ plane.

Figure 4. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): spatial data consolidation maps after ArchaeOBIA processing of the field drawings. |A| Consolidation of data for spatial analysis in the XY plane, expressed as a percentage per square meter of remains with reconstructed (x, y) coordinates. |B| Consolidation of data for spatial analysis in the XYZ plane, expressed as a percentage per square meter of remains with (-, -, z) coordinates reconstructed in (x, y) (Modified from Peudon (2021)).
Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : cartes de consolidation de l’information spatiale suite au traitement ArchéOBIA des relevés de terrain. |A| Consolidation des données spatiales exploitables en analyse spatiale dans le plan xy, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y) reconstituées. |B| Consolidation des données spatiales exploitables en analyse spatiale dans le plan xyz, exprimée en pourcentage de vestiges avec coordonnées (-, — , z) reconstituées en (x, y) (Modifié d’après Peudon (2021)).

Figure 4. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): spatial data consolidation maps after ArchaeOBIA processing of the field drawings. |A| Consolidation of data for spatial analysis in the XY plane, expressed as a percentage per square meter of remains with reconstructed (x, y) coordinates. |B| Consolidation of data for spatial analysis in the XYZ plane, expressed as a percentage per square meter of remains with (-, -, z) coordinates reconstructed in (x, y) (Modified from Peudon (2021)). Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : cartes de consolidation de l’information spatiale suite au traitement ArchéOBIA des relevés de terrain. |A| Consolidation des données spatiales exploitables en analyse spatiale dans le plan xy, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y) reconstituées. |B| Consolidation des données spatiales exploitables en analyse spatiale dans le plan xyz, exprimée en pourcentage de vestiges avec coordonnées (-, — , z) reconstituées en (x, y) (Modifié d’après Peudon (2021)).

35As for the morphometric information, 16 attributes were calculated to consolidate and enrich the descriptive database of the remains. The ArchaeOBIA length and width attributes have enabled us to reconstruct the missing dimensions of 4,996 remains (i.e., 21% of the remains). Moreover, 14 new morphometric attributes were added, building a new dataset for the site of Cagny-l’Épinette. This ArchaeOBIA enrichment, uneven from one area and excavation campaign to another, ranges from 0% to 100% of remains drawn, and therefore described by these new attributes (fig. 5). Fig. 5 shows areas where the drawing protocol was not applied, which explains the zero rates obtained in some square meters. It also highlights areas where the field drawing protocol was applied, but where the percentage of remains is very low or nil. These square meters are indicative of field drawings that must have existed, but are missing from the archives.

Figure 5. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): morphometric data enrichment map after ArchaeOBIA processing of the field drawings, expressed as a percentage per square meter of remains described by the 14 new attributes.
Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : carte d’enrichissement des données morphométriques suite au traitement ArchéOBIA des relevés de terrain, exprimé en pourcentage, par mètre carré, de vestiges décrits par les 14 nouveaux attributs.

Figure 5. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): morphometric data enrichment map after ArchaeOBIA processing of the field drawings, expressed as a percentage per square meter of remains described by the 14 new attributes. Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : carte d’enrichissement des données morphométriques suite au traitement ArchéOBIA des relevés de terrain, exprimé en pourcentage, par mètre carré, de vestiges décrits par les 14 nouveaux attributs.

5.2 | ArchaeOBIA versus classical approach: a cross-validation

36The ArchaeOBIA-generated data (i.e., xy coordinates and length of remains) were compared with the data obtained by the «classic» approach (i.e., field and laboratory data) in order to validate the merging of these two datasets (Peudon 2021).

5.2.1 | Field versus ArchaeOBIA geolocations

37The spatial data produced (xy coordinates) are derived from field measurements (manual and total station) and ArchaeOBIA output reconstructed measurements. It is therefore possible to compare the quality of their planimetric georeferencing. This comparison is applied to remains with both xy data sources (field and ArchaeOBIA) by calculating the absolute difference (Δ) between the two sets of xy coordinates (Peudon 2021). The dataset of comparison comprises 11,375 remains (table 1). However, 1995 proved to be an anomalous year in terms of protocol (see Peudon 2021); these data (1,234 out of 11,375) have therefore been excluded from the analysis.

Table 1.Cagny-l’Épinette — Distribution of the (xy) spatial dataset before and after curation: number of archaeological remains with or without xy coordinates according to the source of the spatial information, i.e., original field measurements (manual or total station) or data reconstructed from field drawings.
Cagny-l’Épinette — Distribution du jeu de données spatiales (x, y) avant et après curation : nombre de vestiges archéologiques avec ou sans coordonnées (x, y) selon la source de l’information spatiale, c’est-à-dire les mesures originales de terrain (manuelles ou station totale) ou les données reconstituées à partir des relevés de terrain.

Table 1.Cagny-l’Épinette — Distribution of the (xy) spatial dataset before and after curation: number of archaeological remains with or without xy coordinates according to the source of the spatial information, i.e., original field measurements (manual or total station) or data reconstructed from field drawings. Cagny-l’Épinette — Distribution du jeu de données spatiales (x, y) avant et après curation : nombre de vestiges archéologiques avec ou sans coordonnées (x, y) selon la source de l’information spatiale, c’est-à-dire les mesures originales de terrain (manuelles ou station totale) ou les données reconstituées à partir des relevés de terrain.

38The Δx and Δy (i.e., deviations from the exactitude of the geolocation) may come from the manual drawing part or from the coordinate recording part. Furthermore, the georeferencing of manual field drawings in GIS is a potential source of additional inaccuracy during the data curation process.

39Over the course of the excavation period, 5 different recording combinations were used, depending on the evolution of field drawing and measurement methods. These combinations all mostly show a variation of a few centimeters in the field (Δx and Δy), equivalent to a few millimeters on manual drawings. In detail, some combinations are a little more consistent than others, i.e., whose Δx and Δy variations are lower (fig. 6). The “Graph paper/Field drawing copy” and “Graph paper/Manual” combinations provide the most consistent spatial information, with respectively 79.4% and 89% of variations less than 5 cm (i.e., less than 5 mm on drawing). The “Graph paper/Total station”, “Photograph/Total station” and “Tracing paper/Total station” combinations show slightly lower rates, with respectively 76.5%, 77.4% and 79.2% of variations less than 5 cm.

Figure 6. Cagny-l’Épinette — Δx and Δy variations according to the combinations of drawing and xy measurement methods applied in the field. Outliers, considered to be human errors, have been removed from the box plots |A, B| for easier reading; they are included in table |C|. “Drawing method/xy measurement method” combinations: 1 — “Graph paper/Field drawing copy”, 2 — “Graph paper/Manual”, 3 — “Graph paper/Total station”, 4 — “Photograph/Total station”, 5 — “Tracing paper/Total station” (Modified from Peudon (2021)).
Cagny-l’Épinette — Variations Δx et Δy selon les combinaisons de méthodes de relevé et de mesures (x, y) appliquées sur le terrain. Les valeurs hors normes, considérées comme des erreurs humaines, ont été enlevées des box plots |A, B| pour une meilleure lecture ; elles sont incluses dans le tableau |C|. Combinaisons « Méthode de relevé/Méthode de mesure des coordonnées (x, y) » : 1 — « Millimétré/Copie relevé », 2 — « Millimétré/Manuelle », 3 — « Millimétré/Tachéomètre », 4 — « Photo/Tachéomètre », 5 — « Calque/Tachéomètre » (Modifié d’après Peudon (2021)).

Figure 6. Cagny-l’Épinette — Δx and Δy variations according to the combinations of drawing and xy measurement methods applied in the field. Outliers, considered to be human errors, have been removed from the box plots |A, B| for easier reading; they are included in table |C|. “Drawing method/xy measurement method” combinations: 1 — “Graph paper/Field drawing copy”, 2 — “Graph paper/Manual”, 3 — “Graph paper/Total station”, 4 — “Photograph/Total station”, 5 — “Tracing paper/Total station” (Modified from Peudon (2021)). Cagny-l’Épinette — Variations Δx et Δy selon les combinaisons de méthodes de relevé et de mesures (x, y) appliquées sur le terrain. Les valeurs hors normes, considérées comme des erreurs humaines, ont été enlevées des box plots |A, B| pour une meilleure lecture ; elles sont incluses dans le tableau |C|. Combinaisons « Méthode de relevé/Méthode de mesure des coordonnées (x, y) » : 1 — « Millimétré/Copie relevé », 2 — « Millimétré/Manuelle », 3 — « Millimétré/Tachéomètre », 4 — « Photo/Tachéomètre », 5 — « Calque/Tachéomètre » (Modifié d’après Peudon (2021)).

40Few remains have a large margin of error. The highest margins of error concern the “Photograph/Total station” (1.1 m (Δx), 1.8 m (Δy)) and “Tracing paper/Total station” (1.4 m (Δx), 0.8 m (Δy)) combinations, compared with the other combinations (1 m (Δx), 0.8–1 m (Δy)).

41The box plots illustrate these distribution differences. They show a more variable Δx and Δy statistical distribution for the combinations involving a total station than for the others.

42These Δx and Δy correspond to the inaccuracy of the measurement tools. The coordinates measured in the field correspond to the target point of the archaeologist on the artifact, whereas eCognition calculates the coordinates from the exact centroid of each image object. Moreover, the pencil stroke (i.e., pencil stroke thickness and position inaccuracy compared to reality) on the drawings (e.g., scaled values of ± 1–2 mm on drawing are equivalent to real values of ± 1–2 cm) adds to these inaccuracies. However, despite these inaccuracies, both the coordinates measured in the field and ArchaeOBIA are for the most part consistent for all five combinations. The inter-combination distinction discussed here therefore relates to a very low distribution variability. Indeed, the inclusion of the extreme values in the box plots makes this variability completely unreadable.

43Overall, there is a better data consistency when only the human agent is involved in recording the various data in the field, i.e., “Graph paper/Field drawing copy” (coordinates read from the drawings) and “Graph paper/Manual” combinations. In the cases of the “Tracing paper/Total station” and “Photograph/Total station” combinations, their lesser consistency can be explained by mirror movements when taking measurements with the total station. It can also be explained by the fact that the photographs are not orthorectified. The “Graph paper/Total station” combination shows a lesser consistency as well. This can be explained by the combination of a strictly manual practice, drawing on graph paper, with a strictly computerized practice, measuring with a total station.

44Finally, the few extreme discrepancies (fig. 6, C) are due to unavoidable human error, especially on a multi-decennial excavation. It is interesting to point out here that this rate of values considered as extreme (i.e., remains with real values of Δx or Δy ≥ 10 cm) remains low in number for four out of five combinations. The introduction of a total station has therefore not reduced the frequency of human error.

45Regardless of the combination, the ArchaeOBIA processing enables us to reconstruct the missing (x, y) coordinates of several hundred of remains with good accuracy. The two sources of spatial information, direct from the field and indirect from ArchaeOBIA, are sufficiently consistent and compatible to be used simultaneously in a robust spatial analysis.

5.2.2 | Morphometry: caliper versus ArchaeOBIA measurements

46The caliper measurements are compared here, for the length of the remains, with the ArchaeOBIA-reconstructed measurements (Peudon 2021):

  • by artifact,

  • by size class of remains.

47The comparison by artifact was applied to remains with both caliper and ArchaeOBIA length measurements, i.e., 8,166 remains (table 2). The length variations (ΔL) can be explained by the manual variability, i.e., output data after caliper measurement and ArchaeOBIA processing input drawing type. Furthermore, the difference in accuracy between the caliper and ArchaeOBIA measurements has a relative importance depending on the dimensions of the remains.

Table 2. Cagny-l’Épinette — Distribution of the morphometric dataset (Length) before and after curation: number of archaeological remains with or without length measurement according to the source of the morphometric information, i.e., original caliper measurements or data reconstructed from field drawings.
Cagny-l’Épinette — Distribution du jeu de données morphométriques (Longueur) avant et après curation : nombre de vestiges archéologiques avec ou sans mesure de la longueur selon la source de l’information morphométrique, c’est-à-dire les mesures originales au pied à coulisse ou les données reconstituées à partir des relevés de terrain.

Table 2. Cagny-l’Épinette — Distribution of the morphometric dataset (Length) before and after curation: number of archaeological remains with or without length measurement according to the source of the morphometric information, i.e., original caliper measurements or data reconstructed from field drawings. Cagny-l’Épinette — Distribution du jeu de données morphométriques (Longueur) avant et après curation : nombre de vestiges archéologiques avec ou sans mesure de la longueur selon la source de l’information morphométrique, c’est-à-dire les mesures originales au pied à coulisse ou les données reconstituées à partir des relevés de terrain.

*Including 599 remains < 20 mm without exact measurement.
**Including 1021 remains < 20 mm without exact measurement.
***Including 1620 remains < 20 mm without exact measurement.

48The variations observed among the drawing methods are in line with the advantages and disadvantages of each type of drawing (fig. 7). Drawing on a photograph, with 93.6% of ΔL ≤ 20 mm (i.e., 2 mm on drawing), shows the best metric match with the caliper measurements. Drawing on tracing paper is a little less accurate (84.4% of ΔL ≤ 20 mm), as the paper opacity affects the perception of the underlying photograph. Finally, the strictly manual drawings on graph paper show more discrepancy between caliper and ArchaeOBIA measurements, with 73.5% of ΔL ≤ 20 mm. The box plots illustrate the more variable statistical distribution of ΔL for drawings on graph paper compared with those on a photograph or tracing paper.

Figure 7. Cagny-l’Épinette — Length variations (ΔL) between caliper and ArchaeOBIA measurements according to field drawing types. Outliers, included in table |B|, have been removed from the box plot |A| for easier reading (Modified from Peudon (2021)).
Cagny-l’Épinette — Variations de longueur (ΔL) entre les mesures manuelles (au pied à coulisse) et ArchéOBIA selon les types de relevé de terrain. Les valeurs hors normes, incluses dans le tableau |B|, ont été enlevées du box plot |A| pour une meilleure lecture (Modifié d’après Peudon (2021)).

Figure 7. Cagny-l’Épinette — Length variations (ΔL) between caliper and ArchaeOBIA measurements according to field drawing types. Outliers, included in table |B|, have been removed from the box plot |A| for easier reading (Modified from Peudon (2021)). Cagny-l’Épinette — Variations de longueur (ΔL) entre les mesures manuelles (au pied à coulisse) et ArchéOBIA selon les types de relevé de terrain. Les valeurs hors normes, incluses dans le tableau |B|, ont été enlevées du box plot |A| pour une meilleure lecture (Modifié d’après Peudon (2021)).

49The variations due to the size of the remains show a progressive increase shared by the latter and the Δ (fig. 8). The small remains (“L ≤ 20 mm” and “20 < L ≤ 50 mm” classes) show the best consistency between caliper and ArchaeOBIA measurements; variations are the smallest (respectively 83.8% and 85.2% of ΔL ≤ 20 mm). Conversely, the rate of ΔL ≤ 20 mm drops for larger remains (down to 28.6% for the “200 < L ≤ 250 mm” class). The impact of the size of the remains on the Δ is therefore to be understood in terms of percentage of the final measurement, as it is sensitive to a size effect of the remains, e.g., a 5 mm discrepancy between caliper and ArchaeOBIA measurements means a relative variation of ± 20% for a remains < 20 mm, and ± 5% for a remains ≥ 10 cm. To assess the value of merging the datasets produced by the two measurement methods, it is therefore necessary to group the remains by size classes, in order to eliminate the size effect.

Figure 8. Cagny-l’Épinette — Length variations (ΔL) between caliper and ArchaeOBIA measurements according to size classes (length) of the remains. Outliers, included in table |B|, have been removed from the box plot |A| for easier reading (Modified from Peudon (2021)).
Cagny-l’Épinette — Variations de longueur (ΔL) entre les mesures manuelles et ArchéOBIA selon les classes de dimensions (longueur) des vestiges. Les valeurs hors normes, incluses dans le tableau |B|, ont été enlevées du box plot |A| pour une meilleure lecture (Modifié d’après Peudon (2021)).

Figure 8. Cagny-l’Épinette — Length variations (ΔL) between caliper and ArchaeOBIA measurements according to size classes (length) of the remains. Outliers, included in table |B|, have been removed from the box plot |A| for easier reading (Modified from Peudon (2021)). Cagny-l’Épinette — Variations de longueur (ΔL) entre les mesures manuelles et ArchéOBIA selon les classes de dimensions (longueur) des vestiges. Les valeurs hors normes, incluses dans le tableau |B|, ont été enlevées du box plot |A| pour une meilleure lecture (Modifié d’après Peudon (2021)).

50The comparison by size classes of the remains from both datasets (caliper versus ArchaeOBIA measurements) is based on the calculation of natural breaks using the method of Jenks (1967). This classification method generates homogeneous classes by minimizing the intra-class variance and maximizing the inter-class variance. It was also used for statistical data mining to define the adequate number of classes.

51Several Jenks classifications were carried out in ArcGIS software (ArcMap 10.7) on three subsets of data, i.e., the 8,166 remains with both caliper and ArchaeOBIA lengths, the 12,813 remains with caliper measurements and the 4,996 remains for which the missing length was reconstructed by ArchaeOBIA (table 2). The last two subsets correspond to the final dataset content used for spatial analysis. This comparison enables us to assess whether the class distribution changes according to the subsets.

52The first subset (8,166 remains) enables us to assess the stability of frequency and of deviation between the two methods by comparing caliper and ArchaeOBIA measurements for the same remains (tables 3 and 4). 6-, 7- or 8-class options are best suited to data classification (table 3). The differences in break values between caliper and ArchaeOBIA measurements are very small, i.e., just a few millimeters. The degree of manual inaccuracy is therefore consistent whatever the method, particularly for the small remains (e.g., pencil stroke thickness). Moreover, the distribution of remains within the different size classes is equivalent to within a few percent (table 4). There is therefore the same frequency of size observation between caliper and ArchaeOBIA measurements. Finally, the results confirm the low impact of the remains without exact measurements (i.e., lithics < 20 mm and very large faunal remains) on the distribution of classes and class populations.

Table 3. Cagny-l’Épinette — Dataset of remains (n = 8,166) with both caliper and ArchaeOBIA length measurements: Comparison of break values of Jenks-generated size classes between the two measurement types. Color-coded according to number of classes: 4- and 5-class classifications (white), 6-, 7- and 8-class classifications (red, blue and yellow respectively).
Cagny-l’Épinette — Corpus des vestiges (n = 8 166) avec mesures de longueur manuelle (au pied à coulisse) et ArchéOBIA : comparaison, entre les deux sources de mesures, des valeurs de rupture des classes de taille obtenues par la méthode de Jenks. Code couleur : classifications à 4- et 5-classes en blanc, classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.

Table 3. Cagny-l’Épinette — Dataset of remains (n = 8,166) with both caliper and ArchaeOBIA length measurements: Comparison of break values of Jenks-generated size classes between the two measurement types. Color-coded according to number of classes: 4- and 5-class classifications (white), 6-, 7- and 8-class classifications (red, blue and yellow respectively). Cagny-l’Épinette — Corpus des vestiges (n = 8 166) avec mesures de longueur manuelle (au pied à coulisse) et ArchéOBIA : comparaison, entre les deux sources de mesures, des valeurs de rupture des classes de taille obtenues par la méthode de Jenks. Code couleur : classifications à 4- et 5-classes en blanc, classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.

Table 4. Cagny-l’Épinette — Dataset of remains (n = 8,166) with both caliper and ArchaeOBIA length measurements: Comparison of the distribution of remains within the Jenks-generated size classes, between the two measurement types. Color-coded according to number of classes: 6-, 7- and 8-class classifications (red, blue and yellow respectively).
Cagny-l’Épinette — Corpus des vestiges (n = 8 166) avec mesures de longueur manuelle (au pied à coulisse) et ArchéOBIA : comparaison, entre les deux sources de mesures, de la distribution des vestiges au sein des classes de dimensions obtenues par la méthode de Jenks. Code couleur : classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.

Table 4. Cagny-l’Épinette — Dataset of remains (n = 8,166) with both caliper and ArchaeOBIA length measurements: Comparison of the distribution of remains within the Jenks-generated size classes, between the two measurement types. Color-coded according to number of classes: 6-, 7- and 8-class classifications (red, blue and yellow respectively). Cagny-l’Épinette — Corpus des vestiges (n = 8 166) avec mesures de longueur manuelle (au pied à coulisse) et ArchéOBIA : comparaison, entre les deux sources de mesures, de la distribution des vestiges au sein des classes de dimensions obtenues par la méthode de Jenks. Code couleur : classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.

53The other two subsets (12,813 and 4,996 remains) enable us to assess the distribution sensitivity of the remains when merging a caliper-measured dataset and an ArchaeOBIA-reconstructed dataset (tables 5 and 6). The results of the 6-, 7- and 8-classes show very similar size classes between the two datasets, with the exception of the last large-size classes (table 5). This confirms that caliper and ArchaeOBIA measurements show similar statistical distributions. Furthermore, the measurement error between the two methods is relatively small for the 6-, 7- or 8-classes. Reconstructing data using ArchaeOBIA therefore does not influence the database accuracy. Finally, the two subsets show similar distributions of the remains within the different size classes, to within a few percent (table 6). Consequently, there is no statistical reason not to merge the two datasets in the same site-wide analysis (Peudon 2021).

Table 5. Cagny-l’Épinette — Merged analysis dataset, i.e., remains with caliper length measurement (n = 12,813) and remains with ArchaeOBIA-reconstructed length measurement (n = 4,996): Comparison of break values of Jenks-generated size classes between the two measurement types. Color-coded according to number of classes: 4- and 5-class classifications (white), 6-, 7- and 8-class classifications (red, blue and yellow respectively) (Modified from Peudon (2021)).
Cagny-l’Épinette — Corpus d’analyse agrégé, i.e., vestiges avec mesures au pied à coulisse (n = 12 813) et vestiges avec mesures reconstituées sous ArchéOBIA (n = 4 996) : comparaison, entre les deux sources de mesures, des valeurs de rupture des classes de taille obtenues par la méthode de Jenks. Code couleur : classifications à 4- et 5-classes en blanc, classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement (Modifié d’après Peudon (2021)).

Table 5. Cagny-l’Épinette — Merged analysis dataset, i.e., remains with caliper length measurement (n = 12,813) and remains with ArchaeOBIA-reconstructed length measurement (n = 4,996): Comparison of break values of Jenks-generated size classes between the two measurement types. Color-coded according to number of classes: 4- and 5-class classifications (white), 6-, 7- and 8-class classifications (red, blue and yellow respectively) (Modified from Peudon (2021)). Cagny-l’Épinette — Corpus d’analyse agrégé, i.e., vestiges avec mesures au pied à coulisse (n = 12 813) et vestiges avec mesures reconstituées sous ArchéOBIA (n = 4 996) : comparaison, entre les deux sources de mesures, des valeurs de rupture des classes de taille obtenues par la méthode de Jenks. Code couleur : classifications à 4- et 5-classes en blanc, classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement (Modifié d’après Peudon (2021)).

Table 6. Cagny-l’Épinette — Merged analysis dataset, i.e., remains with caliper length measurement (n = 12,813) and remains with ArchaeOBIA-reconstructed length measurement (n = 4,996): Comparison of the distribution of remains within the Jenks-generated size classes, between the two measurement types. Color-coded according to number of classes: 6-, 7- and 8-class classifications (red, blue and yellow respectively).
Cagny-l’Épinette — Corpus d’analyse agrégé, i.e., vestiges avec mesures au pied à coulisse (n = 12 813) et vestiges avec mesures reconstituées sous ArchéOBIA (n = 4 996) : comparaison, entre les deux sources de mesures, de la distribution des vestiges au sein des classes de dimensions obtenues par la méthode de Jenks. Code couleur : classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.

Table 6. Cagny-l’Épinette — Merged analysis dataset, i.e., remains with caliper length measurement (n = 12,813) and remains with ArchaeOBIA-reconstructed length measurement (n = 4,996): Comparison of the distribution of remains within the Jenks-generated size classes, between the two measurement types. Color-coded according to number of classes: 6-, 7- and 8-class classifications (red, blue and yellow respectively). Cagny-l’Épinette — Corpus d’analyse agrégé, i.e., vestiges avec mesures au pied à coulisse (n = 12 813) et vestiges avec mesures reconstituées sous ArchéOBIA (n = 4 996) : comparaison, entre les deux sources de mesures, de la distribution des vestiges au sein des classes de dimensions obtenues par la méthode de Jenks. Code couleur : classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.

54Our data curation using ArchaeOBIA on several thousand remains enables us to assess the consistency of data from different methods, and to fill in any data gaps, ensuring that the reconstructed data does not introduce a bias for future statistical analyses.

5.3 | Discussion

5.3.1 | Data integrity

55During the data curation step, we were confronted with Bennett et al.’s (1984) concepts of missing information and missing data. According to these authors, missing data refers to the case where “[…] data are available in the correct form for the purposes of an analysis, but some individual items are missing from the data set in the possession of the analyst.”, while “[…] missing information occurs when data have been or could have been recorded in full but are not available to the analyst. […] The missing information problem therefore refers to cases in which data are available but not in precisely the required form.” At Cagny-l’Épinette, the question of both missing data and missing information, arose for the archaeological remains and their semantic, spatial and morphometric descriptions (Peudon 2021). Consequently, as considering data quality is an essential prerequisite for any study (Bennett et al. 1984; Cheetham and Haigh 1992; Wong and Lee 2005), spatial analyses could only be considered once the data integrity had been assessed.

56In terms of coordinates, 66.5% of remains have a complete set (x, y, z) after data curation. 12.9% were assigned an incomplete set (x, y, -), while less than 0.1% have an incomplete set (-, -, z). A fifth of the remains (20.6%) have no coordinates. In terms of dimensions, 35.4% of remains have a complete set (L, W, THK). 39.5% have an incomplete set (L, W, -), (L, -, -) or (L, -, THK). A quarter of the remains (25.1%) have no dimensions. As for the new ArchaeOBIA morphometric data, it describes 54.9% of the remains, i.e., only those with a field drawing. As in fig. 5 and fig. 9, the integrity of the final data has been quantified and mapped for each archaeological level (see Peudon (2021) for these maps). With such documents, the spatial analysis approach has been adapted to the data accuracy, which has itself been taken into account in the interpretations.

Figure 9. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): Integrity maps of the final information after data curation. |A| Integrity of data for spatial analysis in the XYZ plane, expressed as a percentage per square meter of remains with (x, y, z) coordinates. |B| Integrity of data for spatial analysis in the XY plane, expressed as a percentage per square meter of remains with (x, y, -) or (x, y, z) coordinates.
Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : cartes d’intégrité de l’information finale après curation des données. |A| Intégrité de la donnée spatiale exploitable dans le plan XYZ, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y, z). |B| Intégrité de la donnée spatiale exploitable dans le plan XY, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y, — ) ou (x, y, z).

Figure 9. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): Integrity maps of the final information after data curation. |A| Integrity of data for spatial analysis in the XYZ plane, expressed as a percentage per square meter of remains with (x, y, z) coordinates. |B| Integrity of data for spatial analysis in the XY plane, expressed as a percentage per square meter of remains with (x, y, -) or (x, y, z) coordinates. Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : cartes d’intégrité de l’information finale après curation des données. |A| Intégrité de la donnée spatiale exploitable dans le plan XYZ, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y, z). |B| Intégrité de la donnée spatiale exploitable dans le plan XY, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y, — ) ou (x, y, z).

5.3.2 | Critical assessment and feedback

57The ArchaeOBIA/GIS protocol has proved to be a powerful tool for automating and optimizing the extraction of information from a large set of diversified archives (Peudon 2021). It has been an effective means of data gathering and inter-data linking. Furthermore, developing a classification rule set with eCognition proved to be a relatively straightforward task in the case of pre-digitized and geo-referenced handwritten field drawings. The designed classification rule set can be reused on similar data and, if necessary, edited to fit new datasets. In our case, its application took less than a minute per field drawing. Finally, our protocol demonstrates eCognition usefulness in the automated transformation of raster images into shape files for documents already vectorized using CAD software. And this, while optimizing the ratio between processing time and final data quality.

58Through this approach, we demonstrate that the handwritten field drawings of Cagny-l’Épinette, archives of the geometric and spatial features of archaeological remains, can be used to reconstruct lost data or to obtain data that was not originally recorded. In line with previous studies (Benito-Calvo and de la Torre 2011; Boschian and Saccà 2010; De La Torre and Benito-Calvo 2013), this confirms the status of field drawings as a reliable source of information when exploiting legacy archives or in the case of digital file loss. Our position therefore differs from that of Domínguez-Rodrigo et al. (2012). The latter reject this type of information extraction on the grounds of possible morphometric inaccuracy in the rendering of remains on field drawings and of possible geometric inaccuracy of the modeled features of the remains (e.g., longitudinal axis), biasing the features derived from them (e.g., orientation). In the case of Cagny-l’Épinette, drawings were either rigorously produced in an XY Cartesian coordinate system, or directly generated from photographs. As for the calculation of object features, eCognition returns georeferenced, geometrically accurate features, e.g., georeferenced orientation of the real longitudinal axis (Peudon 2021).

59However, this methodology has some limitations. As the overlapping of several shapes affects the integrity of the spatial and geometric surface information of at least one of them, particular attention must be paid to the non-overlapping of remains, which increases the archive processing time for data curation. Furthermore, although average values can be calculated from the control points indicated on the field drawings (e.g., elevation), not all information can be reconstructed (e.g., plunge, precise elevation).

60Finally, as with other existing protocols, the results of this methodology are affected by errors occurring during the original information recording (detectable through cross-comparison of data) and by the degree of accuracy of the field drawing and its digitization (morphometric variations of a few millimeters cannot be avoided at a 1/10 scale).

6 | Conclusion: a meta-archaeology of Cagny-l’Épinette

61Our approach belongs to the digital humanities applied to the production and curation of archaeological data. It deploys digital processing tools and concepts to provide input for the prehistorian’s thematic analysis. For large datasets, data curation is an essential step in validating the quality of the knowledge produced by the archaeologist. In the context of a long-term excavation, data curation is all the more important as it optimizes the investment in technical and human resources by consolidating technical and thematic knowledge.

62At Cagny-l’Épinette, the evolution of the archives, juxtaposing and hybridizing recording methods over 30 years, has resulted in a complex data accumulation. Our ArchaeOBIA protocol, applied to handwritten field drawings, has enabled us to significantly consolidate and enrich the database of archaeological remains, and to assess its reliability. In doing so, it has enabled us to completely revise the spatial analysis of the remains at the site of Cagny-l’Épinette. With regard to data consolidation, the number of valid and reliable remains has thus been increased by 7.5% (i.e., 1,788 remains) for geolocation and by 21% (i.e., 4,996 remains) for morphometric description. 14 new attributes were also calculated using a standardized, reproducible protocol for 54.9% of the remains (i.e., 13,077 remains). To assess the reliability of the data, our analysis was based on several thousand initial field and laboratory recordings, compared with the data reconstructed after the consolidation step. While this assessment validates the ArchaeOBIA data, the quality of data curation remains dependent on the accuracy and uncertainty of the initial data (handwritten field drawings, caliper measurements, digital measurements, etc.) from the excavation and study of the material.

63Beyond the Cagny-l’Épinette case study, our ArchaeOBIA protocol can also be used on larger datasets. It can also be used in a processing chain mobilizing data from multiple acquisition sources (e.g., handwritten drawings, total station, 3D scanner, drone photogrammetry, drone LiDAR, etc.). In this paper, revisiting the archived data from a different methodological perspective is indeed a proof of concept of meta-archaeology and its relevance to long-term excavations.

Top of page

Bibliography

AGACHE R. 1971 - Nord et Picardie. Gallia Préhistoire 14, 2, p. 271–310.

ANTOINE P. 1990 - Chronostratigraphie et environnement du Paléolithique du bassin de la Somme. Publications du CERP, 2. Centre d’Études et de Recherches Préhistoriques, Université des Sciences et Techniques de Lille Flandres-Artois, Villeneuve d’Ascq, 220 p.

ANTOINE P., TUFFREAU A. 1993 - Contexte stratigraphique, climatique et paléotopographique des occupations acheuléennes de la moyenne terrasse de la Somme. Bulletin de la Société préhistorique française, 90, 4, p. 243–250. https://doi.org/10.3406/bspf.1993.9589

ATICI L., KANSA S.W., LEV-TOV J., KANSA E.C. 2013 - Other people’s data: a demonstration of the imperative of publishing primary data. Journal of Archaeological Method and Theory, 20, p. 663–681. https://doi.org/10.1007/s10816-012-9132-9

AUGUSTE P., LAMOTTE A., LOCHT J.-L., TUFFREAU A. 2005 - Le traitement de la matière première lithique et osseuse au Paléolithique inférieur et Paléolithique moyen dans le Nord de la France : état des recherches récentes. In : Molines N., Moncel M.-H., Monnier J.L. (Éds.), Les premiers peuplements en Europe. Oxford, BAR international series. John and Erica Hedges, p. 419–431.

BAATZ M., HOFFMANN C., WILLHAUCK G. 2008 - Progressing from object-based to object-oriented image analysis. In : Blaschke T., Lang S., Hay G.J. (Eds.), Object-Based Image Analysis: Spatial Concepts for Knowledge-Driven Remote Sensing Applications, Lecture Notes in Geoinformation and Cartography. Springer, Berlin, Heidelberg, p. 29–42. https://doi.org/10.1007/978-3-540-77058-9_2

BAATZ M., SCHÄPE A. 2000 - Multiresolution segmentation: an optimization approach for high quality multi-scale image segmentation. In : Strobl J., Blaschke T., Griesebner G. (Eds.), Beiträge Zum AGIT-Symposium Salzburg 2000, Angewandte Geographische Informationsverarbeitung. Wichmann, Heidelberg, p. 12–23.

BAHAIN J.-J., FALGUÈRES C., LAURENT M. 2001 - Datation par résonance paramagnétique électronique (RPE) de sédiments et par combinaison des méthodes RPE/U-Th de restes paléontologiques (provenant des sites paléolithiques de Cagny). In : Tuffreau A. (Éd.), L’Acheuléen dans la vallée de la Somme et Paléolithique moyen dans le Nord de la France : données récentes. Publications du CERP 6. Centre d’Études et de Recherches Préhistoriques, Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, p. 37–40.

BARGE O., CHATAIGNER C., RÉGAGNON E., GOUX A. 2008 - L’ASPRO : un exemple d’interface cartographique pour la consultation d’un corpus archéologique. In : Djindjian F., Noizet H., Costa L., Pouget F. (Éds.), Webmapping dans les sciences historiques et archéologiques : actes du colloque international (Paris, 3–4 juin 2008). All’Insegna del Giglio, p. 137–151.

BATES M.R. 1993 - Quaternary aminostratigraphy in North-western France. Quaternary Science Reviews, 12, p. 793–809. https://doi.org/10.1016/0277-3791(93)90018-H

BENITO-CALVO A., TORRE (de la) I. 2011 - Analysis of orientation patterns in Olduvai Bed I assemblages using GIS techniques: implications for site formation processes. Journal of Human Evolution, 61, p. 50–60. https://doi.org/10.1016/j.jhevol.2011.02.011

BENNETT R.J., HAINING R.P., GRIFFITH D.A. 1984 - The problem of missing data on spatial surfaces. Annals of the American Association of Geographers, 74, p. 138–156. https://doi.org/10.1111/j.1467-8306.1984.tb01440.x

BERNARD L. 2019 - Ten years of ArkeoGIS: feedback on the challenges and benefits of digitally sharing archeological data. In : Sharing the Experience: Workflows for the Digital Humanities [Dariah-CH Workshop #2], 5–6 Dec 2019, Neuchâtel (Switzerland).

BIRD D., MIRANDA L., VANDER LINDEN M., ROBINSON E., BOCINSKY R.K., NICHOLSON C., CAPRILES J.M., FINLEY J.B., GAYO E.M., GIL A., ALPOIM GUEDES (d’) J., HOGGARTH J.A., KAY A., LOFTUS E., LOMBARDO U., MACKIE M., PALMISANO A., SOLHEIM S., KELLY R.L., FREEMAN J. 2022 - p3k14C, a synthetic global database of archaeological radiocarbon dates. Scientific Data 9, 1, 27 p. https://doi.org/10.1038/s41597-022-01118-7

BLASCHKE T. 2003 - Object-based contextual image classification built on image segmentation. In : Proceedings of the 2003 IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, Greenbelt, 27–28 October 2003. Presented at the IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, IEEE, p. 113–119. https://doi.org/10.1109/WARSD.2003.1295182

BLASCHKE T. 2010 - Object based image analysis for remote sensing. ISPRS Journal of Photogrammetry and Remote Sensing, 65, p. 2–16. https://doi.org/10.1016/j.isprsjprs.2009.06.004

BLASCHKE T., LANG S. 2006 - Object based image analysis for automated information extraction — A synthesis. In : Measuring the Earth II: ASPRS-MAPPS Fall Conference, San Antonio, Texas, November 6–10, 2006. Presented at the Measuring the Earth II: ASPRS Fall Conference.

BLASCHKE T., STROBL J. 2001 - What’s wrong with pixels? Some recent developments interfacing remote sensing and GIS. Zeitschrift für Geoinformationssysteme, p. 12–17.

BOSCHIAN G., SACCÀ D. 2010 - Ambiguities in human and elephant interactions? Stories of bones, sand and water from Castel di Guido (Italy). Quaternary International, 214, p. 3–16. https://doi.org/10.1016/j.quaint.2009.10.016

BOURDIER F., CHALINE J., MUNAUT A.V., PUISSÉGUR J.J. 1974 - Le complexe Mindelien : II — La moyenne terrasse de l’Avre. Quaternaire, 11, p. 168–180. https://doi.org/10.3406/quate.1974.2127

BOUYSSI J.-P., FONTAINE A., ROBERT E., VIALOU D., VILHENA VIALOU A., HOULLIER J.-R., EGELS Y. 2012 - Fressignes, Campement Solutréen Sur Ordinateur. Archeologia e Calcolatori Supplemento, 3, p. 355–369.

BUCHSENSCHUTZ O., CARTEREAU M., LAMBERT G.-N. 1986 - CENTRAR : un système de gestion et de représentation graphique des données archéologiques. Bulletin de la Société préhistorique française, 83, 10, Études et travaux, p. 325–327. https://doi.org/10.3406/bspf.1986.8708

BUCHSENSCHUTZ O., DEBANNE X. 1978 - Quatre années de traitement informatique des données de fouilles à Levroux (Indre). ArcheoSciences, Revue d’Archéométrie, 2, p. 19–27. https://doi.org/10.3406/arsci.1978.1090

CERRILLO-CUENCA E. 2017 - An approach to the automatic surveying of prehistoric barrows through LiDAR. Quaternary International, Staring at the ground: Archaeological surveys as a research tool, 435, p. 135–145. https://doi.org/10.1016/j.quaint.2015.12.099

CHANG Y.-C., SHYU B.-K., WANG J.-S. 1997 - Region-based fractal image compression with quadtree segmentation. In : Presented at 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing, IEEE Computer Society Press, Munich, Germany, p. 3125–3128. https://doi.org/10.1109/ICASSP.1997.595454

CHEETHAM P., HAIGH J.G.B. 1992 - The archaeological database ‒ New relations? In : Lock G., Moffett J. (Eds.), Computer Applications and Quantitative Methods in Archaeology 1991. Oxford, BAR International Series. Tempus Reparatum, p. 7–14.

CHENHALL R.G. 1967 - The description of archaeological data in computer language. American Antiquity, 32, p. 161–167. https://doi.org/10.2307/277901

CHENHALL R.G. 1968 - The impact of computers on archaeological theory: an appraisal and projection. Computers and the Humanities, 3, p. 15–24. https://doi.org/10.1007/BF02395445

CHENHALL R.G. 1971 - The archaeological data bank: a progress report. Computers and the Humanities, 5, p. 159–169. https://doi.org/10.1007/BF02402182

COWGILL G.L. 1967 - Computer applications in archaeology. Computers and the Humanities, 2, p. 17–23. https://doi.org/10.1007/BF02402460

D’ANDREA A., GALLOTTI R., PIPERNO M. 2002 - A GIS intra-site application to the Early Palaeolithic site of Garba IV (Melka Kunture, Ethiopia). In : Raynal J.-P., Albore-Livadie C., Piperno M. (Éds.), Hommes et Volcans — De l’éruption à l’objet. Actes du Symposium 15.2. Les Dossiers de l’Archéo-Logis, p. 93–99.

DAVIS D.S. 2018 - Object-based image analysis: a review of developments and future directions of automated feature detection in landscape archaeology. Archaeological Prospection, 26, p. 155–163. https://doi.org/10.1002/arp.1730

DAVIS D.S. 2020 - Geographic disparity in machine intelligence approaches for archaeological remote sensing research. Remote Sensing 12, p. 921. https://doi.org/10.3390/rs12060921

DAVIS D.S., COOK HALE J.W., HALE N.L., JOHNSTON T.Z., SANGER M.C. 2024 - Bathymetric LiDAR and Semi-Automated Feature Extraction Assist Underwater Archaeological Surveys. Archaeological Prospection, 31, p. 171–186. https://doi.org/10.1002/arp.1939

DAVIS D.S., LIPO C.P., SANGER M.C. 2019 - A comparison of automated object extraction methods for mound and shell-ring identification in coastal South Carolina. Journal of Archaeological Science: Reports, 23, p. 166–177. https://doi.org/10.1016/j.jasrep.2018.10.035

DE GUIO A., MAGNINI L., BETTINESCHI C. 2015 - GeOBIA approaches to remote sensing of fossil landscapes: two case studies from Northern Italy. In : Traviglia A. (Ed.), Across Space and Time: Papers from the 41st Conference on Computer Applications and Quantitative Methods in Archaeology, Perth, 25–28 March 2013, Applications and Quantitative Methods in Archaeology. Presented at the 41st Conference on Computer Applications and Quantitative Methods in Archaeology, Amsterdam University Press, Amsterdam, p. 45–53. https://doi.org/10.5117/9789089647153

DE LAET V., PAULISSEN E., WAELKENS M. 2007 - Methods for the extraction of archaeological features from very high-resolution Ikonos–2 remote sensing imagery, Hisar (southwest Turkey). Journal of Archaeological Science, 34, p. 830–841. https://doi.org/10.1016/j.jas.2006.09.013

DEUSY P. 2003 - Relation lithique-faune du gisement acheuléen de Cagny-l’Épinette. Mise à jour et interprétation des plans du secteur 22N à 26P (1997–2002) — Niveaux I1, I1B et I2. Volume concernant les parcelles 25O, 26O, 24P, 25P et 26P (Mémoire d’initiation à la recherche). Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, France.

DIBBLE H.L. 1987 - Measurement of artifact provenience with an electronic theodolite. Journal of Field Archaeology, 14, p. 249–254.

DIBBLE H.L., CHASE P.G., McPHERRON S.P., TUFFREAU A. 1997 - Testing the reality of a “living floor” with archaeological data. American Antiquity, 62, p. 629–651. https://doi.org/10.2307/281882

DIBBLE H.L., McPHERRON S.P. 1988 - On the computerization of archaeological projects. Journal of Field Archaeology, 15, p. 431–440. https://doi.org/10.1179/jfa.1988.15.4.431

DOMINGUEZ-RODRIGO M., BUNN H.T., PICKERING T.R., MABULLA A.Z.P., MUSIBA C.M., BAQUEDANO E., ASHLEY G.M., DIEZ-MARTIN F., SANTONJA M., URIBELLARREA D., BARBA R., YRAVEDRA J., BARBONI D., ARRIAZA C., GIDNA A. 2012 - Autochthony and orientation patterns in Olduvai Bed I: a re-examination of the status of post-depositional biasing of archaeological assemblages from FLK North (FLKN). Journal of Archaeological Science, 39, p. 2116–2127. https://doi.org/10.1016/j.jas.2012.02.027

DRUBAY J.B. 2003 - Relation lithique-faune du gisement acheuléen de Cagny-l’Épinette. Mise à jour et interprétation des plans du secteur 22N à 26P (1997–2002) — Niveau I1B/I2. Volume concernant les parcelles 26M, 26N, 25M, 25N, 24M (Mémoire d’initiation à la recherche). Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, France.

ESTEVA M., TRELOGAN J., RABINOWITZ A., WALLING D., PIPKIN S. 2010 - From the site to long-term preservation: a reflexive system to manage and archive digital archaeological data. Society for Imaging Science and Technology. Archiving 2010. Proceedings of the Archiving Conference, 7, p. 1–6. https://doi.org/10.2352/issn.2168-3204.2010.7.1.art00001

FINKEL R.A., BENTLEY J.L. 1974 - Quad trees a data structure for retrieval on composite keys. Acta Informatica, 4, p. 1–9. https://doi.org/10.1007/BF00288933

FREELAND T., HEUNG B., BURLEY D.V., CLARK G., KNUDBY A. 2016 - Automated feature extraction for prospection and analysis of monumental earthworks from aerial LiDAR in the Kingdom of Tonga. Journal of Archaeological Science, 69, p. 64–74. https://doi.org/10.1016/j.jas.2016.04.011

FRONZA V. 2016 - OpenArcheo2. An archaeological knowledge base. In : Basso P., Caravale A., Grossi P. (Eds.), ArcheoFOSS: Free, libre and open source software e open format nei processi di ricerca archeologica: Atti del IX Workshop (Verona, 19–20 giugno 2014). All’Insegna del Giglio, p. 254–262.

FRONZA V., NARDINI A., VALENTI M. 2003 - An integrated information system for archaeological data management: latest developments. In : Doerr M., Sarris A. (Eds.), Proceedings of CAA 2002: The Digital Heritage of Archaeology. Presented at the CAA 2002, Athens, Greece: Hellenic Ministry of Culture, p. 147–153.

FU K.S., MUI J.K. 1981 - A survey on image segmentation. Pattern Recognition, 13, 1, p. 3–16. https://doi.org/10.1016/0031-3203(81)90028-5

GALLOTTI R. 2004 - Analisi spaziali e metodologie computazionali per un approccio cognitivo ai modelli di frequentazione antropica del giacimento di Isernia La Pineta (Molise, Italia) (Ph.D. Thesis). Ferrara: Ferrara University.

GALLOTTI R., MOHIB A., GRAOUI M.E., SBIHI-ALAOUI F.Z., RAYNAL J.-P. 2011 - GIS and Intra-Site Spatial Analyses: An Integrated Approach for Recording and Analyzing the Fossil Deposits at Casablanca Prehistoric Sites (Morocco). Journal of Geographic Information System, 3, 4, p. 373–381. https://doi.org/10.4236/jgis.2011.34036

GALLOTTI R., PIPERNO M. 2003 - Recent activities of the Italian Archaeological Mission at Melka Kunture (Ethiopia): the Open-Air Museum Project and the GIS application to the study of the Oldowan sites. In : Martínez Moreno J., Mora Torcal R., de la Torre Sáinz I. (Eds.), Oldowan: Rather More than Smashing Stones. First Hominid Technology Workshop, Bellaterra, December 2001, Treballs d’arqueologia 9. Universitat autonoma de Barcelona, Centre d’estudis del patrimoni arqueologic de la prehistoria. Bellaterra, p. 37–75.

GARDIN J.-C. 1971 - Archaeology and computers: new perspectives. International Social Science, 23, 2, p. 189–203.

GARDIN J.-C. 1974 - Les projets de banques de données archéologiques. Problèmes méthodologiques, technologiques et institutionnels. In : Borillo M., Gardin J.-C. (Éds.), Banques de données archéologiques, Marseille 12–14 juin 1972, Colloques nationaux du Centre national de la recherche scientifique. Éditions du Centre national de la recherche scientifique, Paris, p. 15–26.

GARDIN J.-C., GARELLI P. 1961 - Étude des établissements assyriens en Cappadoce par ordinateurs. Annales. Économies, Sociétés et Civilisations, 16, 5, p. 837–876. https://doi.org/10.3406/ahess.1961.420758

GINOUVÈS R. 1971 - Archéographie, archéométrie, archéologie. Pour une informatique de l’archéologie gréco-romaine. Revue Archéologique, p. 93–126.

GINOUVÈS R., GUIMIER-SORBETS A.M. 1978 - La constitution des données en archéologie classique : recherches et expériences en vue de la préparation de bases de données. Éditions du CNRS, Paris, 161 p., 17 fig.

GOODSON K.J. 1989 - Shape Information in an artefact Database. In : Rahtz S., Richards J. (Eds.), Computer Applications and Quantitative Methods in Archaeology 1989. Oxford, BAR International Series, 548, p. 349–361.

GORTON D., SHEN R., VEMURI N.S., FAN W., FOX E.A. 2006 - ETANA-GIS: GIS for archaeological digital libraries. In : Marchionini G., Nelson M.L., Marshall C.C. (Eds.), JCDL 2006: Proceedings of the Sixth ACM/IEEE-CS Joint Conference on Digital Libraries: Opening Information Horizons, Chapel Hill, NC, USA, June 11–15, 2006. Presented at the JCDL06: Joint Conference on Digital Libraries 2006, Association for Computing Machinery, New York, NY, USA, p. 379. https://doi.org/10.1145/1141753.1141875

GRACE R., GRAHAM I.D.G., NEWCORNER M.H. 1985 - The quantification of microwear polishes. World Archaeology, 17, p. 112–120. https://doi.org/10.1080/00438243.1985.9979954

HAESAERTS P., BALESCU S., DUPUIS C., VAN VLIET B. 1984 - Contribution à la stratigraphie des gisements paléolithiques de Cagny (Somme). Cahiers de Géographie Physique, Travaux du Laboratoire de Géomorphologie et d’Étude du Quaternaire, Villeneuve d’Ascq, 5, p. 77–94.

HAESAERTS P., DUPUIS C. 1986 - Contribution à la stratigraphie des nappes alluviales de la Somme et de l’Avre dans la région d’Amiens. Bulletin de l’association française pour l’étude quaternaire, 26, p. 171–186.

HAY G.J., CASTILLA G. 2006 - Object-Based Image Analysis: strengths, weaknesses, opportunities and threats (SWOT). In : Lang S., Blaschke T., Schöpfer E. (Eds.). Bridging Remote Sensing and GIS: Proceedings of the 1st International Conference on Object-Based Image Analysis (OBIA 2006), July 4–5, 2006, Salzburg University, Austria. Presented at the Bridging Remote Sensing and GIS: 1st International Conference on Object-based Image Analysis (OBIA 2006), ISPRS Archives.

HAY G.J., CASTILLA G. 2008 - Geographic Object-Based Image Analysis (GEOBIA): a new name for a new discipline. In : Blaschke T., Lang S., Hay G.J. (Eds.), Object-Based Image Analysis: Spatial Concepts for Knowledge-Driven Remote Sensing Applications, Lecture Notes in Geoinformation and Cartography. Springer, Berlin, Heidelberg, p. 75–89. https://doi.org/10.1007/978-3-540-77058-9_4

HAY G.J., CASTILLA G., WULDER M.A., RUIZ J.R. 2005 - An automated object-based approach for the multiscale image segmentation of forest scenes. International Journal of Applied Earth Observations and Geoinformation, 7, 4, p. 339–359. https://doi.org/10.1016/j.jag.2005.06.005

HAY G.J., MARCEAU D.J., DUBÉ P., BOUCHARD A. 2001 - A multi-scale framework for landscape analysis: object-specific analysis and upscaling. Landscape Ecology, 16, p. 471–490. https://doi.org/10.1023/A:1013101931793

HAY G.J., NIEMANN K.O., GOODENOUGH D.G. 1997 - Spatial thresholds, image-objects, and upscaling: a multiscale evaluation. Remote Sensing Environment, 62, 1, p. 1–19. https://doi.org/10.1016/S0034-4257(97)81622-7

HAY G.J., NIEMANN K.O., McLEAN G.F. 1996 - An object-specific image-texture analysis of H-resolution forest imagery. Remote Sensing. Environment, 55, 2, p. 108–122. https://doi.org/10.1016/0034-4257(95)00189-1

HODDER I. 1997 - “Always momentary, fluid and flexible”: towards a reflexive excavation methodology. Antiquity, 71, p. 691–700. https://doi.org/10.1017/S0003598X00085410

HOFMANN P., MARSCHALLINGER R., UNTERWURZACHER M., ZOBL F. 2013 - Marble provenance designation with Object Based Image Analysis: state-of-the-art rock fabric characterization from petrographic micrographs. Austrian Journal of Earth Science, 106, 2, p. 40–49.

HOULLIER J.-R., ARNOUX T. 2001 - Informatisation d’une fouille : réalisation et déploiement du logiciel d’acquisition de données « FrAcTool » (Fressignes Acquisition Tool). Paleo, 13, p. 251–264. https://doi.org/10.4000/paleo.1096

HUGGETT J., ROSS S. 2004 - Introduction. Internet Archaeo-logy. https://doi.org/10.11141/ia.15.13

HUNTER G.M., STEIGLITZ K. 1979 - Operations on Images Using Quad Trees. IEEE Transactions on Pattern Analysis Machine Intelligence. PAMI–1, p. 145–153. https://doi.org/10.1109/TPAMI.1979.4766900

HUSI P., RODIER X. 2011 - ArSol: an archaeological data processing system. In : Jerem E., Redő F., Szeverényi V. (Eds.), On the Road to Reconstructing the Past: Computer Applications and Quantitative Methods in Archaeology (CAA), Proceedings of the 36th International Conference, Budapest, April 2–6, 2008. Presented at the CAA 2008, Archaeolingua Foundation, Budapest, p. 278–284.

INOMATA T., PINZÓN F., RANCHOS J.L., HARAGUCHI T., NASU H., FERNANDEZ-DIAZ J.C., AOYAMA K., YONENOBU H. 2017 - Archaeological application of airborne LiDAR with object-based vegetation classification and visualization techniques at the lowland Maya site of Ceibal, Guatemala. Remote Sensing, 9. https://doi.org/10.3390/rs9060563

JAHJAH M., ULIVIERI C., INVERNIZZI A., PARAPETTI R. 2007 - Archaeological remote sensing application pre-post war situation of Babylon archaeological site — Iraq. Acta Astronautica, 61, p. 121–130. https://doi.org/10.1016/j.actaastro.2007.01.034

JENKS G.F. 1967 - The data model concept in statistical mapping. International Yearbook of Cartography, 7, p. 186–190.

KANDEL A.W., SOMMER C., KANAEVA Z., BOLUS M., BRUCH A.A., GROTH C., HAIDLE M.N., HERTLER C., HEß J., MALINA M., MÄRKER M., HOCHSCHILD V., MOSBRUGGER V., SCHRENK F., CONARD N.J. 2023 - The ROCEEH Out of Africa Database (ROAD): a large-scale research database serves as an indispensable tool for human evolutionary studies. PLoS ONE, 18, e0289513. https://doi.org/10.1371/journal.pone.0289513

KINTIGH K. 2006 - The promise and challenge of archaeological data integration. American Antiquity, 71, p. 567–578. https://doi.org/10.1017/S0002731600039810

KULASEKARAN S., TRELOGAN J., ESTEVA M., JOHNSON M. 2014 - Metadata integration for an archaeology collection architecture. In : Moen W., Rushing A. (Eds.), Metadata Intersections: Bridging the Archipelago of Cultural Memory, 2014 Proceedings of the International Conference on Dublin Core and Metadata Applications. Presented at the International Conference on Dublin Core and Metadata Applications, Dublin Core Metadata Initiative, p. 53–63.

LAMOTTE A. 1999 - L’apport des remontages dans la compréhension des méthodes de débitage et de façonnage des gisements acheuléens de la Somme : les exemples de la Ferme de l’Épinette et de l’Épinette à Cagny (Somme, France). Bulletin de la Société préhistorique française, 96, 2, p. 117–131. https://doi.org/10.3406/bspf.1999.10936

LAMOTTE A. 2001 - Analyse morpho-fonctionnelle et métrique des bifaces des séries de la séquence fluviatile (séries I0, I1, I1A, I1B/I2) du gisement acheuléen de Cagny-l’Épinette (Somme, France). In : Cliquet D. (Éd.), Les industries à outils bifaciaux du Paléolithique moyen d’Europe occidentale, Actes de la table-ronde internationale de Caen, 14–15 oct. 1999. Liège, ERAUL, 98, p. 21–28.

LAMOTTE A. 2012 - Le Paléolithique inférieur et moyen des bassins-versants de la Somme et de la Saône : aspects cognitifs de l’environnement pétrographique et considérations techno-typologique (Thèse d’Habilitation à Diriger des Recherches). Université de Lille 1, Sciences et Technologies, Villeneuve d’Ascq, France.

LAMOTTE A., FABRE, J. 2007 - Approvisionnement et circulation du silex des sites du Paléolithique inférieur de Cagny-la-Garenne, Cagny-l’Épinette et de la Ferme de l’Épinette (Bassin de l’Avre, Somme, France). In : Moncel M.-H., Moigne A.-M., Arzarello M., Peretto C. (Éds.), Aires d’approvisionnement en matières premières et aires d’approvisionnement en ressources alimentaires : Approche intégrée des comportements, Actes du XVe Congrès Mondial (Lisbonne, 4–9 Septembre 2006), 5, Session WS23. Oxford, BAR International Series. Archaeopress, p. 47–59.

LAMOTTE A., MASSON É. 2016 - Arché-OBIA : un concept d’analyse quantitative d’images numériques appliqué aux bifaces du gisement de Gouzeaucourt (Nord, FR). Notae Praehistoricae, 36, p. 121–130.

LAMOTTE A., TUFFREAU A. 2001 - Les industries lithiques de la séquence fluviatile fine de Cagny-l’Épinette (Somme). In : Tuffreau A. (Éd.), L’Acheuléen dans la vallée de la Somme et Paléolithique moyen dans le Nord de la France : données récentes, Publications du CERP 6. Centre d’Études et de Recherches Préhistoriques, Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, p. 113–134.

LAMOTTE A., TUFFREAU A. 2016 - Acheulean of the Somme basin (France): assessment of lithic changes during MIS 12 to 9. Quaternary International, 409, p. 54–72. https://doi.org/10.1016/j.quaint.2015.08.058

LANG S., BLASCHKE T. 2006 - Bridging Remote Sensing and GIS — What are the main supportive pillars? In : Lang S., Blaschke T., Schöpfer E. (Eds.), Bridging Remote Sensing and GIS: Proceedings of the 1st International Conference on Object-Based Image Analysis (OBIA 2006), July 4–5, 2006, Salzburg University, Austria. Presented at the Bridging Remote Sensing and GIS: 1st International Conference on Object-based Image Analysis (OBIA 2006), ISPRS Archives.

LAURENT M. 1993 - Datation par résonance de spin électronique (ESR) de quartz de formations quaternaires. Comparaison avec le paléomagnétisme (Thèse de doctorat en Terre, Océan, Espace). Muséum national d’Histoire naturelle, Paris.

LAURENT M., FALGUÈRES C., BAHAIN J.J., ROUSSEAU L., VAN VLIET-LANOÉ B. 1998 - ESR dating of quartz extracted from quaternary and neogene sediments: method, potential and actual limits. Quaternary Science Reviews, 17, p. 1057–1062. https://doi.org/10.1016/S0277-3791(97)00101-7

LAURENT M., FALGUÈRES C., BAHAIN J.-J., YOKOYAMA Y. 1994 - Géochronologie du système de terrasses fluviatiles quaternaires du bassin de la Somme par datation RPE sur quartz, déséquilibres des familles de l’uranium et magnétostratigraphie. Comptes Rendus Hebdomadaires de l’Académie des Sciences, 318, p. 521–526.

LE GOFF E., MARLET O., RODIER X., CURET S., HUSI P. 2015 - Interoperability of the ArSol (Archives du Sol) database based on the CIDOC-CRM ontology. In : Giligny F., Djindjian F., Costa L., Moscati P., Robert S. (Eds.), CAA2014, 21st Century Archaeology: Concepts, Methods and Tools: Proceedings of the 42nd Annual Conference on Computer Applications and Quantitative Methods in Archaeology. Presented at the CAA, Archaeopress, Oxford, p. 179–186.

LEFEVRE B. 2003 - Relation lithique-faune du gisement acheuléen de Cagny-l’Épinette. Mise à jour et interprétation des plans du secteur 22N à 26P (1997–2002) — Niveau I1/I1B/I2. Volume concernant les parcelles 22O ; 23O ; 24O ; 22P ; 23P (Mémoire d’initiation à la recherche). Université des Sciences et Technologies de Lille.

LÉOPOLD J. 1989 - Débitage et outillage sur éclat des industries acheuléennes de Cagny-l’Épinette et de Cagny-Cimetière (Somme). In : Tuffreau A. (Éd.), Paléolithique et Mésolithique du Nord de la France. Nouvelles recherches I, Publications du CERP 1. Centre d’Études et de Recherches Préhistoriques, Université des Sciences et Techniques de Lille Flandres-Artois, Villeneuve d’Ascq, p. 9–18.

LÉOPOLD J., 1993. Les industries lithiques de Cagny-L’Épinette, Cagny-Cimetière et Cagny-la-Garenne dans le contexte Nord-Ouest européen (Thèse de doctorat). Muséum national d’Histoire naturelle, Paris.

LÉOPOLD J. 1997 - Les industries lithiques de Cagny-L’Épinette, Cagny-Cimetière et Cagny-la-Garenne dans le contexte Nord-Ouest européen. L’Anthropologie, 101, p. 639–669.

LORIOT B., FOUGEROLLE Y., SESTIER C., SEULIN R. 2007 - 3D acquisition and modeling for flint artefacts analysis. In : Fotakis C., Pezzati L., Salimbeni R. (Eds.), Proceedings Volume 6618, O3A: Optics for Arts, Architecture, and Archaeology. Presented at the Optical Metrology, 17–21 June 2007, Munich, Germany, SPIE, p. 66180G. https://doi.org/10.1117/12.726149

MAGNINI L., BETTINESCHI C. 2019 - Theory and practice for an object-based approach in archaeological remote sensing. Journal of Archaeological Science, 107, p. 10–22. https://doi.org/10.1016/j.jas.2019.04.005

MAGNINI L., POZZI-ESCOT D., OSHIRO J., ANGELES R., APA M.I.P., VENTURA G. 2024 -Effects of the Architectural Layout of the Sanctuary of Pachacamac (2nd–16th Century CE, Peru) on the Exposure to Rain, Wind, and Solar Radiation from the Morphometric Analysis of Digital Surface Models. Remote Sensing 16, p. 1848. https://doi.org/10.3390/rs16111848

MARLET O., BAUDOIN B., BERNARD L., BÉZARD L., BOISSAT R., BUARD P.-Y., HALCZUK A., HIVERT F., HÉLÈNE J., NOUVEL B., RODIER X., ROUSSET M., SCHOLTUS L. 2022 - Le Livre Blanc du consortium Mémoires des Archéologues et des Sites Archéologiques : guide des bonnes pratiques numériques en archéologie (report). Consortium MASA.

MASSON É., LAMOTTE A. 2018 - Arché-OBIA : vestiges lithiques au Paléolithique et exploitation quantifiée de leurs images numériques. CRAHN-SNEPH, Hors-série 1, p. 153–165.

MATTON P. 2007 - Étude du matériel faunique 2005–2006 du gisement acheuléen de Cagny-l’Épinette : structure de la population et répartition spatiale (Mémoire de Master 1 en Environnement). Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, France.

McKEAGUE P., CORNS A., SHAW R. 2012 - Developing a spatial data infrastructure for archaeological and
built heritage. International Journal of Spatial Data Infrastructures Research, p. 38–65. https://doi.org/10.2902/1725-0463.2012.07.art3

McKEAGUE P., JONES R. 2008 - The National Database for Scotland: evolution of the digital resource. In : Posluschny A., Lambers K., Herzog I. (Eds.), Layers of Perception: Proceedings of the 35th International Conference on Computer Applications and Quantitative Methods in Archaeology (CAA), Berlin, Germany, April 2–6, 2007, Kolloquien Zur Vor- Und Frühgeschichte. Presented at the International Conference on Computer Applications and Quantitative Methods in Archaeology, Dr. Rudolf Habelt Verlag, Bonn, p. 186–193.

McKEAGUE P., Veer (van’t) R., HUVILA I., MOREAU A., VERHAGEN P., BERNARD L., COOPER A., GREEN C., VAN MANEN N. 2019 - Mapping our heritage: towards a sustainable future for digital spatial information and technologies in european archaeological heritage management. Journal of Computers Applications in Archaeology, 2, p. 89–104. https://doi.org/10.5334/jcaa.23

McMANAMON F., KINTIGH K. 2010 - Digital Antiquity: transforming archaeological data into knowledge. SAA Archaeological Record, 10, p. 37–40. https://doi.org/10.6067/XCV8QF8S60

McPHERRON S.P. 1991 - Digitalisation d’images et analyse des matériaux lithiques. In : Tuffreau A. (Éd.), Paléolithique et Mésolithique du Nord de la France. Nouvelles recherches II, Publications du CERP 3. Centre d’Études et de Recherches Préhistoriques, Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, p. 21–26.

McVICAR J. 1986 - Using microcomputers in archaeology: some comments and suggestions. In : Webb E. (Ed.), Computer Applications in Archaeology 1985: Proceedings of the Conference on Quantitative Methods, Institute of Archaeology, London, March 29–30, 1985. Presented at the Computer Applications in Archaeology 1985, Institute of Archaeology, University of London, London, p. 102–108.

MOHLEHLI M.G., ADAM E., SCHOEMAN M.H. 2023 - The potential for Lidar using support vector machine (SYM) to detect archaeological stone-walled structures in Khutwaneng, Bokoni. South African Archaeological Bulletin, 78, p. 33–42. https://www.jstor.org/stable/27250346

MOIGNE A.-M. 1988 - L’étude des vestiges osseux du gisement acheuléen de Cagny-l’Épinette (Somme). Revue Archéologique de Picardie, 1, p. 69–71. https://doi.org/10.3406/pica.1988.1576

MOIGNE A.-M. 1989 - Cagny-l’Épinette : la grande faune. In : Tuffreau A. (Éd.), Livret-guide de l’excursion dans la vallée de la Somme, dans le cadre du colloque « L’Acheuléen dans l’Ouest de l’Europe (Saint-Riquier, 6–10 juin, 1989) », p. 82–87.

MOIGNE A.-M., VALENSI P., AUGUSTE P., GARCIA-SOLANO J., TUFFREAU A., LAMOTTE A., BARROSO C., MONCEL M.-H. 2016 - Bone retouchers from Lower Palaeolithic sites: Terra Amata, Orgnac 3, Cagny-l’Épinette and Cueva del Angel. Quaternary International, 409, p. 195–212. https://doi.org/10.1016/j.quaint.2015.06.059

MURRAY D. 2004 - National inventories: from catalogues to computers. Internet Archaeology https://doi.org/10.11141/ia.15.5

NICCOLUCCI F., RICHARDS J.D. 2019 - ARIADNE and ARIADNEplus. In : Richards J.D., Niccolucci F. (Eds.), The ARIADNE Impact. Archaeolingua, Budapest, p. 7–25. https://doi.org/10.5281/ZENODO.3476712

PAL N.R., PAL S.K. 1993 - A review on image segmentation techniques. Pattern Recognition, 26, p. 1277–1294. https://doi.org/10.1016/0031-3203(93)90135-J

PERROLAS G., NIKNEJAD M., RIBEIRO R., BERNARDINO A. 2022 - Scalable Fire and Smoke Segmentation from Aerial Images Using Convolutional Neural Networks and Quad-Tree Search. Sensors, 22, 1701. https://doi.org/10.3390/s22051701

PEUDON F. 2013 - Le gisement acheuléen de Cagny-l’Épinette de la quatrième terrasse alluviale de la vallée de la Somme (Picardie, France) : contribution à la mise en application d’une méthodologie d’analyse spatiale sur les niveaux archéologiques de la séquence fluviatile (Mémoire de Master 2). Université de Bourgogne, Dijon.

PEUDON F. 2021 - Données archéologiques et archéologie des données : palethnologie du gisement acheuléen de Cagny-l’Épinette (Somme, France) (Thèse de doctorat en Histoire, Civilisations, Archéologie des Mondes antiques). Université de Lille, École Doctorale Sciences de l’Homme et de la Société, Villeneuve d’Ascq, France.

PEUDON F., MASSON É., AUGUSTE P., LAMOTTE A., MOIGNE A.-M., TUFFREAU A. 2021 - Cagny-l’Épinette (Somme Valley, France), thirty years of mixed data: potential and limits. In : Djindjian F., Moscati P. (Eds.), Big Data and Archaeology: Proceedings of the XVIII UISPP World Congress (4–9 June 2018, Paris, France), Volume 15, Session III–1. Presented at the XVIII UISPP World Congress (4–9 June 2018, Paris, France), Archaeopress, Oxford, p. 48–61.

PIROT F., IRIE BI IRIE F., HOBINAVALONA RATSIHOARANA H. 2008 - Implementation of Geoportal in open environment for the community of the Human Sciences and Society. In : Djindjian F., Noizet H., Costa L., Pouget F. (Éds.), Webmapping dans les Sciences historiques et archéologiques : Actes Du Colloque International (Paris, 3–4 Juin 2008). All’Insegna del Giglio, p. 223–229.

PREGESBAUER M., TRINKS I., NEUBAUER W. 2014 - An object oriented approach to automatic classification of archaeological features in magnetic prospection data. Near Surface Geophysics, 12, p. 651–656. https://doi.org/10.3997/1873–0604.2014014

PREWITT J.M.S. 1970 - Object enhancement and extraction. In : Lipkin B.S. (Ed.), Picture Processing and Psychopictorics. Academic Press, New York, p. 75–149.

RABINOWITZ A., SHAW R., BUCHANAN S., GOLDEN P., KANSA E. 2016 - Making sense of the ways we make sense of the past: the PeriodO project. Bulletin of the Institute of Classical studies, 59, p. 42–55. https://doi.org/10.1111/j.2041-5370.2016.12037.x

RAVINDRANATHAN U., SHEN R., GONÇALVES M.A., FAN W., FOX E.A., FLANAGAN J.W. 2004 - ETANA-DL: a digital library for integrated handling of heterogeneous archaeological data. In : Chen H., Christel M., Lim E.-P. (Eds.), JCDL 2004: Proceedings of the Fourth ACM/IEEE-CS Joint Conference on Digital Libraries: Global Reach and Diverse Impact, Tuscon, Arizona, June 7–11, 2004. Presented at the JCDL04: ACM/IEEE Joint Conference on Digital Libraries, Tuscon, Arizona, June 7–11, 2004, Association for Computing Machinery, New York, NY, USA, p. 76–77. https://doi.org/10.1145/996350.996370

RICHARDS J.D. 1997 - Preservation and re-use of digital data: the role of the Archaeology Data Service. Antiquity, 71, p. 1057–1059. https://doi.org/10.1017/S0003598X00086014

RICHARDS J.D. 2004 - Online archives. Internet Archaeology. https://doi.org/10.11141/ia.15.7

ROBIOLLE-GAUTIER C. 1992 - Approche ethnozoologique du site de Cagny-l’Épinette (Somme) — Niveau I1 (Mémoire de D.E.A.). Laboratoire de Géomorphologie et d’Étude du Quaternaire, Université Lille 1, Villeneuve d’Ascq, France.

RYAN N. 1992 - Beyond the relational database: managing the variety and complexity of archaeological data. In : Lock G.R., Moffett J. (Eds.), Computer Applications and Quantitative Methods in Archaeology 1991, British Archaeological Reports International Series. Presented at the Computer Applications in Archaeology Conference, Tempus Reparatum, Oxford, p. 1–6.

RYAN N. 2004 - Databases. Internet Archaeology https://doi.org/10.11141/ia.15.8

SALATIN J.-M. 2007 - La redocumentarisation, un défi pour les sciences de l’information. Études de communication, 30, p. 13–23. https://doi.org/10.4000/edc.428

SAMET H. 1979 - Quadtree Structures for Region Processing. In : Baumann L.S. (Ed.), Proceedings of the Image Understanding Workshop, Held at Los Angeles, California November 7–8, 1979. Presented at the Image Understanding Workshop, Science Applications, Arlington, Va., États-Unis d’Amérique, p. 36–41.

SANTAGATA C., RAYNAL J.-P., GUADELLI J.-L., FERNANDES P., FIORE I. 2007 - SIG intra-site et analyse de la distribution spatiale des vestiges. In : Saint-Anne I (Sinzelles, Polignac, Haute-Loire). Le Paléolithique moyen de l’unité J1, Les dossiers de l’Archéo-Logis. Laussonne, p. 199–258.

SĂRĂSAN A., ARDELEAN A.-C., BĂLĂTRIE A., WEHRHEIM R., TABALDIEV K., AKMATOV K. 2020 - Mapping burial mounds based on UAV-derived data in the Suusamyr Plateau, Kyrgyzstan. Journal of Archaeological Science, 123, 105251. https://doi.org/10.1016/j.jas.2020.105251

SEVARA C., PREGESBAUER M., DONEUS M., VERHOEVEN G., TRINKS I. 2016 - Pixel versus object — A comparison of strategies for the semi-automated mapping of archaeological features using airborne laser scanning data. Journal of Archaeological Science: Reports 5, p. 485–498. https://doi.org/10.1016/j.jasrep.2015.12.023

SHUSTERMAN E., FEDER M. 1994 - Image compression via improved quadtree decomposition algorithms. IEEE Transactions on Image Process, 3, 2, p. 207–215. https://doi.org/10.1109/83.277901

SPELLEMAEKER E. 2003 - Relation lithique-faune du gisement acheuléen de Cagny-l’Épinette. Mise à jour et interprétation des plans du secteur 22N à 26P (1997–2002) — Niveaux I1B ; I1B/I2 ; I2. Volume concernant les parcelles 22N ; 23N ; 24N ; 22M ; 23M (Mémoire d’initiation à la recherche). Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, France.

SUN S., MU L., FENG R., CHEN Y., HAN W. 2024 - Quadtree decomposition-based Deep learning method for multi-scale coastline extraction with high-resolution remote sensing imagery. Science of Remote Sensing. 9, 100112. https://doi.org/10.1016/j.srs.2023.100112

TENNANT E.W. 2007 - A sample geodatabase structure for managing archaeological data and resources with ArcGIS. Technical Briefs in Historical Archaeology, 2, p. 12–23.

TORRE (de la) I., BENITO-CALVO A. 2013 - Application of GIS methods to retrieve orientation patterns from imagery; a case study from Beds I and II, Olduvai Gorge (Tanzania). Journal of Archaeological Science, 40, p. 2446–2457. https://doi.org/10.1016/j.jas.2013.01.004

TRIER Ø.D., LARSEN S.Ø., SOLBERG R. 2009 - Automatic detection of circular structures in high-resolution satellite images of agricultural land. Archaeological Prospection, 16, p. 1–15. https://doi.org/10.1002/arp.339

TRIMBLE 2016. eCognition® Developer 9.2. User Guide. Trimble Documentation, München, Germany.

TUFFREAU A. 1989 - Cagny-l’Épinette : le cadre stratigraphique. In : Tuffreau A. (Éd.), Livret-guide de l’excursion dans la vallée de la Somme, dans le cadre du colloque « L’Acheuléen dans l’Ouest de l’Europe (Saint-Riquier, 6–10 juin, 1989) », p. 72–74.

TUFFREAU A. 2005 - L’Acheuléen. Revue archéologique de la Picardie, 3, p. 25–27. https://doi.org/10.3406/pica.2005.2453

TUFFREAU A., AMELOOT-VAN DER HEIJDEN N. 1991 - L’Acheuléen de la séquence fluviatile fine de Cagny-l’Épinette (Somme). In : Tuffreau A. (Éd.), Paléolithique et Mésolithique du Nord de la France. Nouvelles recherches II, Publications du CERP 3. Centre d’Études et de Recherches Préhistoriques, Université des Sciences et Technologies de Lille, Villeneuve d’Ascq, p. 27–34.

TUFFREAU A., ANTOINE P., CHASE P.G., DIBBLE H.L., ELLWOOD B.В., KOLFSCHOTEN (van) T., LAMOTTE A., LAURENT M., McPHERRON S.P., MOIGNE A.-M., MUNAUT A.-V. 1995 - Le gisement acheuléen de Cagny-l’Épinette (Somme). Bulletin de la Société préhistorique française, 92, 2, p. 169–192. https://doi.org/10.3406/bspf.1995.10004

TUFFREAU A., BOUCHET J.-P., MOIGNE A.-M., MUNAUT A.-V. 1986 - Les niveaux acheuléens de la moyenne terrasse du bassin de la Somme à Cagny-l’Épinette (Somme). L’Anthropologie, 90, p. 9–27.

TUFFREAU A., LAMOTTE A., GOVAL É. 2008 - Les industries acheuléennes de la France septentrionale. L’Anthropologie, 112, p. 104–139. https://doi.org/10.1016/j.anthro.2008.01.003

TUFFREAU A., LAMOTTE A., MARCY J. 1997 - Land-use and site function in Acheulean complexes of the Somme Valley. World Archaeology, 29, p. 225–241. https://doi.org/10.1080/00438243.1997.9980375

TUFFREAU A., MARCY J. 2002 - Acquisition et exploitation numérique de données archéologiques. In : Miskovsky J.-C. (Éd.), Géologie de la Préhistoire : méthodes, techniques, applications. Association pour l’étude de l’environnement géologique de la Préhistoire, Paris, p. 1375–1389.

TUFFREAU A., MUNAUT A.-V., PUISSÉGUR J.-J., SOMMÉ J. 1982 - Stratigraphie et environnement des industries acheuléennes de la moyenne terrasse du bassin de la Somme (région d’Amiens). Quaternaire, 19, p. 73–81. https://doi.org/10.3406/quate.1982.1423

VAN VLIET-LANOË B. 1989 - Observations paléo-pédologiques sur les gisements de Cagny-la Garenne, Cagny-l’Épinette et Saint-Acheul. In : Tuffreau A. (Éd.), Livret-guide de l’excursion dans la vallée de la Somme, dans le cadre du colloque « L’Acheuléen dans l’Ouest de l’Europe (Saint-Riquier, 6–10 juin, 1989) », p. 95–99.

VERHAGEN P., DRĂGUT L. 2012 - Object-based landform delineation and classification from DEMs for archaeological predictive mapping. Journal of Archaeological Science, 39, p. 698–703. https://doi.org/10.1016/j.jas.2011.11.001

WALLIS J.C., ROLANDO E., BORGMAN C.L. 2013 - If we share data, will anyone use them? Data sharing and reuse in the long tail of science and technology. PLoS ONE, 8, e67332. https://doi.org/10.1371/journal.pone.0067332

WILLHAUCK G. 2000 - Comparison of object oriented classification techniques and standard image analysis for the use of change detection between SPOT multispectral satellite images and aerial photos. In : Beek K.J., Molenaar M., Timmerman J. (Eds.), ISPRS Archives. Presented at the XIXth ISPRS Congress, July 16–23, 2000 Amsterdam, The Netherlands, p. 35–42.

WILLMES C., BECKER D., VERHEUL J., YENER Y., ZICKEL M., BOLTEN A., BUBENZER O., BARETH G. 2017 - PaleoMaps: SDI for open paleoenvironmental GIS data. International Journal of Spatial Data Infrastructures Research, 12, p. 39–61. https://doi.org/10.2902/1725-0463.2017.12.art3

WITHARANA C., OUIMET W.B., JOHNSON K.M. 2018 - Using LiDAR and GEOBIA for automated extraction of eighteenth–
late nineteenth century relict charcoal hearths in southern New England. GIScience Remote Sensing, 55, p. 183–204. https://doi.org/10.1080/15481603.2018.1431356

WONG D.W.S., LEE J. 2005 - Statistical analysis of geographic information with ArcView GIS and ArcGIS, Fully revue & updated edition. Wiley, Hoboken, N.J., USA.

ZHANG Y.J. 1996 - A survey on evaluation methods for image segmentation. Pattern Recognition. 29, p. 1335–1346. https://doi.org/10.1016/0031-3203(95)00169-7

Top of page

Appendix

Données supplémentaires

Les suppléments en ligne sont accessibles ici :

https://nakala.fr/​collection/​10.34847/​nkl.d4fa6dzf

Top of page

List of illustrations

Title Figure 1. |A| Location of the site of Cagny-l’Épinette (black star annotated “Ep”) in the Somme basin. |B| Density map of the inventoried remains (levels H to J), expressed as a number of remains per square meter (Map of France: Modified from IGN 2016 Open source; DEM data: RGE ALTI® 5 m (© IGN), modified; Hydrographic network map: BD CARTHAGE® (© IGN), modified; Density Map: F.l Peudon). |A| Localisation du site de Cagny-l’Épinette (étoile noire annotée « Ep ») dans le bassin de la Somme. |B| Carte de densité des vestiges inventoriés (niveaux H à J), exprimée en nombre de vestiges par mètre carré (Carte de France : IGN 2016 Open source, modifiée ; Données MNT : RGE ALTI® 5 m (IGN), modifié ; Carte du réseau hydrographique : BD CARTHAGE® (© IGN), modifiée).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-1.jpg
File image/jpeg, 616k
Title Figure 2. Cagny-l’Épinette — Processing protocol applied to the field drawings (left) and the descriptive data (right) of the archaeological remains, with the software used for each step. See description in text (Modified from Peudon (2021)). Cagny-l’Épinette Protocole de traitement appliqué aux relevés de terrain (à gauche) et aux données descriptives (à droite) des vestiges archéologiques, avec les logiciels sollicités à chaque étape. Voir description dans le texte (Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-2.jpg
File image/jpeg, 212k
Title Figure 3. Cagny-l’Épinette — ArchaeOBIA processing of the digitized versions of the field drawings; example of the area [20–21/Q-S]. |A| Photographs, |B| Field drawing; |C| Rasterized vector data (GeoTIFF format), superimposed remains in dashed line are processed on a separated vector layer. |D to G| — Steps of the classification rule set applied in eCognition: |D| Initial segmentation, |E| Classification of the “Fauna”, “Lithic” and “Background” objects and merging of the “Fauna” and “Lithic” objects, |F| Merge of the “Background” object, |G| Export in shapefile format with an attribute table (|A| A. Tuffreau, |B| G. Leroy, |C to G| Fl. Peudon; Modified from Peudon (2021)). Cagny-l’Épinette — Traitement ArchéOBIA des vectorisations des relevés de terrain ; exemple du secteur [20-21/Q-S]. |A| Photographies ; |B| Relevé de terrain ; |C| Données vectorielles rastérisées (format GéoTIFF), les vestiges superposés en tireté sont traités sur une autre couche vectorielle ; |D à G| Étapes de la règle de traitement appliquée dans le logiciel eCognition : |D| Segmentation initiale, |E| Classification des objets « Faune », « Lithique » et « Fond » et refaçonnement des objets « Faune » et « Lithique », |F| Refaçonnement de l’objet « Fond », |G| Export au format de fichier de formes avec sa table attributaire(|A| A. Tuffreau, |B| G. Leroy, |C à G| F. Peudon ; Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-3.jpg
File image/jpeg, 1.0M
Title Figure 4. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): spatial data consolidation maps after ArchaeOBIA processing of the field drawings. |A| Consolidation of data for spatial analysis in the XY plane, expressed as a percentage per square meter of remains with reconstructed (x, y) coordinates. |B| Consolidation of data for spatial analysis in the XYZ plane, expressed as a percentage per square meter of remains with (-, -, z) coordinates reconstructed in (x, y) (Modified from Peudon (2021)). Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : cartes de consolidation de l’information spatiale suite au traitement ArchéOBIA des relevés de terrain. |A| Consolidation des données spatiales exploitables en analyse spatiale dans le plan xy, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y) reconstituées. |B| Consolidation des données spatiales exploitables en analyse spatiale dans le plan xyz, exprimée en pourcentage de vestiges avec coordonnées (-, — , z) reconstituées en (x, y) (Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-4.jpg
File image/jpeg, 296k
Title Figure 5. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): morphometric data enrichment map after ArchaeOBIA processing of the field drawings, expressed as a percentage per square meter of remains described by the 14 new attributes. Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : carte d’enrichissement des données morphométriques suite au traitement ArchéOBIA des relevés de terrain, exprimé en pourcentage, par mètre carré, de vestiges décrits par les 14 nouveaux attributs.
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-5.jpg
File image/jpeg, 488k
Title Table 1.Cagny-l’Épinette — Distribution of the (xy) spatial dataset before and after curation: number of archaeological remains with or without xy coordinates according to the source of the spatial information, i.e., original field measurements (manual or total station) or data reconstructed from field drawings. Cagny-l’Épinette — Distribution du jeu de données spatiales (x, y) avant et après curation : nombre de vestiges archéologiques avec ou sans coordonnées (x, y) selon la source de l’information spatiale, c’est-à-dire les mesures originales de terrain (manuelles ou station totale) ou les données reconstituées à partir des relevés de terrain.
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-6.jpg
File image/jpeg, 80k
Title Figure 6. Cagny-l’Épinette — Δx and Δy variations according to the combinations of drawing and xy measurement methods applied in the field. Outliers, considered to be human errors, have been removed from the box plots |A, B| for easier reading; they are included in table |C|. “Drawing method/xy measurement method” combinations: 1 — “Graph paper/Field drawing copy”, 2 — “Graph paper/Manual”, 3 — “Graph paper/Total station”, 4 — “Photograph/Total station”, 5 — “Tracing paper/Total station” (Modified from Peudon (2021)). Cagny-l’Épinette — Variations Δx et Δy selon les combinaisons de méthodes de relevé et de mesures (x, y) appliquées sur le terrain. Les valeurs hors normes, considérées comme des erreurs humaines, ont été enlevées des box plots |A, B| pour une meilleure lecture ; elles sont incluses dans le tableau |C|. Combinaisons « Méthode de relevé/Méthode de mesure des coordonnées (x, y) » : 1 — « Millimétré/Copie relevé », 2 — « Millimétré/Manuelle », 3 — « Millimétré/Tachéomètre », 4 — « Photo/Tachéomètre », 5 — « Calque/Tachéomètre » (Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-7.jpg
File image/jpeg, 244k
Title Table 2. Cagny-l’Épinette — Distribution of the morphometric dataset (Length) before and after curation: number of archaeological remains with or without length measurement according to the source of the morphometric information, i.e., original caliper measurements or data reconstructed from field drawings. Cagny-l’Épinette — Distribution du jeu de données morphométriques (Longueur) avant et après curation : nombre de vestiges archéologiques avec ou sans mesure de la longueur selon la source de l’information morphométrique, c’est-à-dire les mesures originales au pied à coulisse ou les données reconstituées à partir des relevés de terrain.
Caption *Including 599 remains < 20 mm without exact measurement. **Including 1021 remains < 20 mm without exact measurement. ***Including 1620 remains < 20 mm without exact measurement.
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-8.jpg
File image/jpeg, 88k
Title Figure 7. Cagny-l’Épinette — Length variations (ΔL) between caliper and ArchaeOBIA measurements according to field drawing types. Outliers, included in table |B|, have been removed from the box plot |A| for easier reading (Modified from Peudon (2021)). Cagny-l’Épinette — Variations de longueur (ΔL) entre les mesures manuelles (au pied à coulisse) et ArchéOBIA selon les types de relevé de terrain. Les valeurs hors normes, incluses dans le tableau |B|, ont été enlevées du box plot |A| pour une meilleure lecture (Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-9.jpg
File image/jpeg, 156k
Title Figure 8. Cagny-l’Épinette — Length variations (ΔL) between caliper and ArchaeOBIA measurements according to size classes (length) of the remains. Outliers, included in table |B|, have been removed from the box plot |A| for easier reading (Modified from Peudon (2021)). Cagny-l’Épinette — Variations de longueur (ΔL) entre les mesures manuelles et ArchéOBIA selon les classes de dimensions (longueur) des vestiges. Les valeurs hors normes, incluses dans le tableau |B|, ont été enlevées du box plot |A| pour une meilleure lecture (Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-10.jpg
File image/jpeg, 296k
Title Table 3. Cagny-l’Épinette — Dataset of remains (n = 8,166) with both caliper and ArchaeOBIA length measurements: Comparison of break values of Jenks-generated size classes between the two measurement types. Color-coded according to number of classes: 4- and 5-class classifications (white), 6-, 7- and 8-class classifications (red, blue and yellow respectively). Cagny-l’Épinette — Corpus des vestiges (n = 8 166) avec mesures de longueur manuelle (au pied à coulisse) et ArchéOBIA : comparaison, entre les deux sources de mesures, des valeurs de rupture des classes de taille obtenues par la méthode de Jenks. Code couleur : classifications à 4- et 5-classes en blanc, classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-11.jpg
File image/jpeg, 116k
Title Table 4. Cagny-l’Épinette — Dataset of remains (n = 8,166) with both caliper and ArchaeOBIA length measurements: Comparison of the distribution of remains within the Jenks-generated size classes, between the two measurement types. Color-coded according to number of classes: 6-, 7- and 8-class classifications (red, blue and yellow respectively). Cagny-l’Épinette — Corpus des vestiges (n = 8 166) avec mesures de longueur manuelle (au pied à coulisse) et ArchéOBIA : comparaison, entre les deux sources de mesures, de la distribution des vestiges au sein des classes de dimensions obtenues par la méthode de Jenks. Code couleur : classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-12.jpg
File image/jpeg, 156k
Title Table 5. Cagny-l’Épinette — Merged analysis dataset, i.e., remains with caliper length measurement (n = 12,813) and remains with ArchaeOBIA-reconstructed length measurement (n = 4,996): Comparison of break values of Jenks-generated size classes between the two measurement types. Color-coded according to number of classes: 4- and 5-class classifications (white), 6-, 7- and 8-class classifications (red, blue and yellow respectively) (Modified from Peudon (2021)). Cagny-l’Épinette — Corpus d’analyse agrégé, i.e., vestiges avec mesures au pied à coulisse (n = 12 813) et vestiges avec mesures reconstituées sous ArchéOBIA (n = 4 996) : comparaison, entre les deux sources de mesures, des valeurs de rupture des classes de taille obtenues par la méthode de Jenks. Code couleur : classifications à 4- et 5-classes en blanc, classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement (Modifié d’après Peudon (2021)).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-13.jpg
File image/jpeg, 120k
Title Table 6. Cagny-l’Épinette — Merged analysis dataset, i.e., remains with caliper length measurement (n = 12,813) and remains with ArchaeOBIA-reconstructed length measurement (n = 4,996): Comparison of the distribution of remains within the Jenks-generated size classes, between the two measurement types. Color-coded according to number of classes: 6-, 7- and 8-class classifications (red, blue and yellow respectively). Cagny-l’Épinette — Corpus d’analyse agrégé, i.e., vestiges avec mesures au pied à coulisse (n = 12 813) et vestiges avec mesures reconstituées sous ArchéOBIA (n = 4 996) : comparaison, entre les deux sources de mesures, de la distribution des vestiges au sein des classes de dimensions obtenues par la méthode de Jenks. Code couleur : classifications à 6-, 7- et 8-classes en rouge, bleu et jaune respectivement.
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-14.jpg
File image/jpeg, 156k
Title Figure 9. Cagny-l’Épinette — Inventoried dataset (levels H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J and J): Integrity maps of the final information after data curation. |A| Integrity of data for spatial analysis in the XYZ plane, expressed as a percentage per square meter of remains with (x, y, z) coordinates. |B| Integrity of data for spatial analysis in the XY plane, expressed as a percentage per square meter of remains with (x, y, -) or (x, y, z) coordinates. Cagny-l’Épinette — Jeu de données recensées (niveaux H, H1, Hx, I, I0, I1, I1A/IB, I1B, I2, I/J et J) : cartes d’intégrité de l’information finale après curation des données. |A| Intégrité de la donnée spatiale exploitable dans le plan XYZ, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y, z). |B| Intégrité de la donnée spatiale exploitable dans le plan XY, exprimée en pourcentage, par mètre carré, de vestiges avec coordonnées (x, y, — ) ou (x, y, z).
URL http://journals.openedition.org/paleo/docannexe/image/10459/img-15.jpg
File image/jpeg, 276k
Top of page

References

Electronic reference

Floriane Peudon and Éric Masson, “The ArchaeOBIA concept applied to data curation for the Acheulean site of Cagny-l’Épinette (Somme Valley, France)”PALEO [Online], 35 | 2025, Online since 15 March 2026, connection on 16 May 2026. URL: http://journals.openedition.org/paleo/10459; DOI: https://doi.org/10.4000/15znh

Top of page

About the authors

Floriane Peudon

Univ. Lille, CNRS, Ministère de la Culture, UMR 8164 — HALMA — Histoire Archéologie Littérature des Mondes Anciens, F–59000 Lille, France. floriane.peudon[at]univ-lille.fr

Éric Masson

Univ. Lille, Univ. Littoral Côte d’Opale, ULR 4477 — TVES — Territoires Villes Environnement & Société, F–59000 Lille, France. eric.masson[at]univ-lille.fr

Top of page

Copyright

CC-BY-NC-ND-4.0

The text only may be used under licence CC BY-NC-ND 4.0. All other elements (illustrations, imported files) may be subject to specific use terms.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search