Digital archaeology today: achievements and challenges

Korobov Dmitry Sergeevich ORCID: 0000-0002-9571-0405 Doctor of History Doctor of Historical Sciences, Head of the Department of Theory and Methodology of the Institute of Archaeology RAS, Professor of RAS 117292, Russia, Moscow, Ulyanova str., 19 dkorobov@mail.ru

Introduction

The gradual penetration of digital technologies into archaeological science has been going on for decades. The appearance of the specialized conference "Computer Applications and Quantitative Methods in Archaeology (CAA)", which first met at the University of Birmingham in the UK in 1973, can be considered a certain milestone on this long journey ^[1]. Since that moment, 50 years ago, the use of computer technology in archaeological research has taken shape as a special direction. Specialized journals have appeared, for example, “Archeologia e Calcolatori”, published in Italy for more than thirty years ^[2], which is accompanied by applications in various European languages ^[3], or the recently emerged “Journal of Computer Applications in Archaeology (JCAA)” ^[4].

Throughout its more than half a century of development, the direction in question has passed several stages. Thus, one of the pioneers of the use of computers in archaeology, the French researcher F. Jinjian highlights the following steps on this path:

– natural science methods in archaeology (1955-1975): semiotics, paleolandscapes, archeometry;

– multidimensional data analysis (1975-1995): the development of quantitative methods in archaeology;

– geographic information systems (1995-2010): from the study of the archaeological monument to the study of archaeological landscapes;

– multi-agent systems (since 1995): a view of ancient societies as a complex system;

– 3D archaeology (since 2010): the realization of the dream of a comprehensive field archaeology ^{[5, p. 13]}.

In this short review, it is impossible to cover all these areas and to fully highlight the history of their development. I see my task in identifying the most basic trends that are observed in modern digital archaeology, and to focus attention on the problems and difficulties that exist in it. The term "Digital Archaeology" used is interpreted by me broadly as a synonym for "Computational Archaeology", while there is also a narrower application of this term as the use of exclusively three-dimensional modeling in archaeological research. In the latter case, it merges with the term "Virtual Archaeology", which is currently gaining great popularity, under the cap of which, as a rule, the creation of three-dimensional images of archaeological landscapes, objects and finds is hidden (see, for example, ^[6]).

Thus, in this review I will touch upon the current state of the use of geographic information systems, digital archives and databases, three-dimensional modeling by photogrammetry and laser scanning methods, as well as the first steps in the application of artificial intelligence in archaeology.

 Geographic information systems

The history of the use of geographic information systems (GIS) in archaeological research begins in the mid-1980s, when, after the advent of personal computers and specialized software (we are talking, first of all, about the ArcInfo program), the first examples of the use of GIS for solving archaeological problems were published ^{[7, p. 16]}. Since then, the bibliography of publications on this topic has grown dramatically, and already in the mid-1990s it numbered more than three hundred titles ^{[8, p. 5]}. Textbooks on the use of GIS in archaeology have appeared ^{[9, 10]}, including in Russian ^{[7, 11]}.

Already in the early 1990s, three main directions of GIS use by archaeologists took shape:

1) protection of archaeological heritage and predictive modeling;

2) GIS application for historical reconstructions;

3) the use of GIS in landscape archaeology ^[12].

It is possible to distinguish other narrower trends in the geoinformation support of archaeological research, but it seems to me that even today these three listed areas fully reflect the main trends in the use of GIS by archaeologists. It seems interesting to trace which of the directions is leading at the present stage. In a recently published article, Fernando Menendez-Marsh and co-authors analyzed over 570 open access publications published in 1990-2022 on the use of GIS in archaeological research and included in the SCOPUS scientific citation database ^[13]. The authors noted that most of the publications made between 1990 and 2010 concerned issues of cultural heritage management (CRM – Cultural Resource Management), whereas in the last 12 years the number of publications devoted to the use of spatial analysis of archaeological data has increased ^{[13, p. 45, 46]}. As before, the main number of publications is devoted to mapping and spatial information management. However, there is an obvious growing interest in various spatial analysis procedures, among which predictive modeling, visibility and density analysis, and three-dimensional GIS come to the fore ^{[13, p. 46, fig. 6]}. 

Based on the generalizations made in the cited article, the share of GIS published on the Internet (WEB-GIS) does not exceed 2% among the published works of 1990-2022 ^{[13, fig. 6]}. Meanwhile, it is necessary to name this direction as the most noticeable trend of recent years. The number of portals open to users with archaeological information mapped in GIS has increased significantly in recent years. As a good example of foreign works of this kind, we can cite a web map of the settlements of Great Britain and Ireland, including data on 4147 monuments ^{[14, 15]}. Similar works are available in our country. It is worth mentioning the portal "Country of Cities", which combines information about the settlements of Volga Bulgaria, which was created at Kazan Federal University ^{[16, 17]}. An interesting example was the geoportal about the archaeological monuments of the Turo-Pyshminsky interfluve ^[18]. Finally, the Institute of Archaeology of the Russian Academy of Sciences has recently launched a project of the portal of the archaeological map of the Russian Federation, which is based on the constantly updated information in the GIS "Archaeological Monuments of Russia" (APR) ^[19]. In the latter case, the data on the location of archaeological monuments are distorted in order to preserve the archaeological heritage – the map in the public domain should not serve as a tool in the looting of ancient settlements and burial grounds by modern looters. However, the created web resource gives an excellent idea of the richness of the archaeological heritage of our country - at the moment it includes almost 53 thousand archaeological sites of all types, epochs and cultures. Work on filling the database continues, and in the future we will be able to obtain a national-scale GIS in the foreseeable future, including all information about the archaeological antiquities of Russia ^{[20, 21]}.

Another example of a domestic GIS, designed for a very complete description of archaeological sites and their subsequent automatic mapping, was created by St. A. Vasiliev at the IIMC RAS in the early 2000s. The Archaeological Information System (AIS) "Archeographer" was originally created as a database with a detailed description system and having a dynamic connection with GIS MapInfo, in which monuments were mapped as point objects ^[22]. Since 2017, work has been underway on an updated version of the program, which has been developed in the form of a server GIS based on QGIS and includes more than 25 linked tables describing survey sites, field work, Open sheets, pits, excavations, monuments and reporting documentation, numbering more than 250 descriptive fields ^[23]. At the moment, this AGIS is actively used in security archeology, mainly on the territory of St. Petersburg and the Leningrad region.

 Digital archives and databases

The creation of large databases combining tens of thousands of records and open to users is another modern trend in digital archaeology. The development of systems for describing archaeological material and rules for its formalization for subsequent analysis using computer systems is the basis for the emergence of computer archaeology. Its roots go back to the pre-computer era, when a special analytical direction emerged within the framework of "New" (or procedural) archaeology, the purpose of which was to develop clear analytical procedures that allow the archaeological material to be structured according to a system of signs, and further mathematical analysis was supposed to make it possible to identify the most significant of them and on their basis to isolate archaeological cultures which, in turn, would become the basis for further historical interpretations. This is how the famous "analytical machine", dreamed of by David Clark, who is considered the leader of this field of procedural archaeology at the turn of the 1960s - 1970s, was supposed to work ^{[24, pp. 219-242]}.  

Subsequently, after the advent of personal computers, the development of databases with systems for describing a variety of archaeological material became a routine procedure ^{[25, 26]}, which no longer requires serious theoretical justification. The emergence of new computer languages and formats allow combining heterogeneous information accumulated in previous years into one common array. The result of this process was the explosive growth of open databases of archaeological, anthropological, archeogenetic, archeometric and archeobotanical material, which allow any user to access the accumulated vast amounts of information. Thus, BigData is being introduced into everyday archaeological practice ^[27], although there is still little data in the public domain.

An excellent example for the creation of open databases with archaeological information in the future can serve as open-access paleogenetic archives, for example, wide-genomic data of more than 10 thousand ancient and 10 thousand modern individuals analyzed by 1.2 million single nucleotide polymorphisms (SNP), which are actively used by all modern paleogeneticists (The Allen Ancient DNA Resource (AADR) - ^[28]). There is a similar trend in the research of bioarchaeologists – a review of existing domestic data banks of paleoanthropological information and their some foreign analogues was made recently by M.V. Dobrovolskaya ^[29]. There are examples of open databases for using the results of isotopic analysis of the composition of non-ferrous metals, posted on the FLAME (Flow of Ancient Metal across Eurasia) website Oxford University (https://flame.arch.ox.ac.uk /). A similar project was developed there on the generalization of archeobotanical information about agricultural cultures of medieval England of the Anglo-Saxon period (FeedSax - Feeding Anglo-Saxon England) with a digital archive of photographic images of macro-botanic remains available for review (https://portal.sds.ox.ac.uk/feedsax ).     

Perhaps, as one of the few examples of such large data banks with archaeological materials, which was available to users some time ago, the “Montelius” database can serve, which in 2015 included more than 960 thousand scanned images of archaeological objects from Central Europe, mainly the I millennium AD ^[30].. The developer of this software product, Peter Stadler (Vienna Museum of Natural History), provided the database with a software package for automatic serialization, correspondent analysis and mapping of finds and conducted statistical processing of data on antiquities of the Avar period ^[31]. Subsequently, this program was used to study cultural monuments of linear ceramics ^[32]. There is a website for this software product, which, unfortunately, has not been supported since 2017. (https://www.winserion.org/index.html ).

On the other hand, there are numerous examples of publicly available digital archives of individual archaeological expeditions and projects. In this series, one of the most striking is the digital archive of the American School of Classical Archaeology in Athens, on whose website all digitized archaeological materials of the excavations of the Athenian agora since 1931 are published, including field diaries, drawings, photographs of finds (more than 370 thousand digitized images in total), which are combined into a single system using hyperlinks (https://agora.ascsa.net/research ). A similar archive has been prepared for research materials in Corinth; it includes in total more than 260 thousand digital images (https://corinth.ascsa.net/research ).

An attempt to create a unified system for describing archaeological sites in the process of field work at the modern level also requires the involvement of a variety of digital technologies. Usually, large archaeological expeditions are engaged in the development of such systems, within the framework of which it is required to involve a large number of excavation managers or sites of one large excavation, or the heads of individual detachments. Special software products are being created that are installed on mobile phones or tablets and include a database with developed forms describing the monument, its site, cultural layer, features of stratigraphy, typical objects and archaeological material. Such systems are equipped with graphic editors and make it possible to take photographs or small digital drawings, which are included in an electronic field diary in the form of images. Archaeological finds can receive an automatic unique number using the created QR code and thus be immediately included in the database in the form of a field inventory. The experience of such developments is summarized in a recently published collection of articles with open access ^[33].

 Three-dimensional modeling by photogrammetry

The authors of the articles of the above-mentioned collection emphasize that a revolutionary breakthrough in the field of digital fixation of the archaeological excavation process was the use of a photogrammetric method for creating three-dimensional computer models ^[34]. These models can be used at different spatial levels – starting from vast areas of the landscape and ending with individual finds. For different scale of modeling, different technologies of digital fixation are used, using unmanned aerial vehicles (UAVs) ^[35] or manual photographic shooting using a tripod and without it ^[36]. In the last few years, there has been a huge interest in these technological techniques, which are gradually beginning to replace the traditional procedures of archaeological fixation with the help of drawings and drawings. The intensification of this process is especially noticeable when fixing complex architectural structures, for example, mounds with stone structures, or during large-scale archaeological security excavations, when a high speed of field research is required to fulfill a contract with construction organizations ^[36]. There are examples of the introduction of the procedure of digital fixation by photogrammetry and in expeditions working with purely scientific purposes ^[37-39].

This method of implementing digital modeling of archaeological monuments and objects has gained wide popularity, primarily due to its accessibility from the point of view of finance and the relative simplicity of the software. However, the highest quality results are achieved by specialized teams, where professionals are engaged in such shooting and subsequent processing. As examples of the work of domestic specialists in this field, we should mention the rich experience accumulated by the RSSDA laboratory under the leadership of Yu.M. Svoysky, which can be found on the website of this organization (http://www.rssda.su ). 

 Laser scanning

The use of photogrammetry for modeling archaeological objects and finds gradually began to compete with another method of three–dimensional modeling in archaeology - laser scanning, which has been gaining popularity since the early 2000s. The creation of high-precision three-dimensional models of art objects and architectural objects, such as, for example, the pyramids at Giza or the Roman Colosseum, gave impetus to the introduction of this technology in field archaeological research ^{[40, 41]}. There are excellent examples of the use of laser scanning technology in the excavation of archaeological sites ^[42], to a greater extent these technologies are used to create models of archaeological objects ^[43]. This direction of digital archaeology has become especially popular in the study of stone inventory from sites and locations of the Paleolithic era, since the use of photogrammetric technology for photographing objects made of stone has its limitations (primarily due to the illumination and glare of flint or obsidian tools when using a flash). Advanced foreign experience in this area ^[44] is successfully implemented and applied by domestic archaeologists, for example, in the laboratory "Digit" of IAET SB RAS ^[45] (https://archaeology.nsc.ru/proekty/cifra /).

Despite the higher accuracy of computer models created using laser scanning, compared with the photogrammetric method, the first technology is currently much inferior to the second in terms of breadth of application. The main reasons for this phenomenon are the rather high cost of laser scanning equipment and the difficulty in mastering software for its subsequent computer processing. The conducted special comparative studies of the two technologies on the same archaeological sites have shown an absolute advantage of photogrammetry in comparison with laser scanning in terms of time, relative simplicity and accessibility of execution ^[46]. Nevertheless, laser scanning continues to develop successfully, including in the archaeological field of application, especially when it comes to LiDAR technology, which also in its own way has become a revolutionary breakthrough in the search and fixation of archaeological objects hidden under forests. The appearance of this technology in the early 2000s and the first steps towards its application by archaeologists ^[47-49] in recent years have been replaced by a fairly widespread practice due to the appearance of lidar devices on UAVs ^[50]. It is worth noting that, for example, at the Sixth International Conference "Archeology and Geoinformatics" recently held in Moscow, six reports were made on the use of LiDAR technology for searching archaeological sites and fixing archaeological objects from the air ^[51]. Interested readers can get acquainted with them in the format of video presentations on the website of the Institute of Archaeology of the Russian Academy of Sciences ^[52]. Of particular note are the new features provided by Apple mobile phones and tablets equipped with lidar scanning devices, which allows the widespread use of this technology in manual mode ^[53].

 Artificial intelligence

Finally, a few words should be said about the most dynamically developing area related to artificial intelligence. The analysis of large data arrays using deep machine learning of artificial neural networks is becoming an urgent need in a variety of fields. There are the first examples of such works in archaeology. For example, a team of Spanish archaeologists has recently developed machine learning algorithms for automatic recognition and mapping of mounds on old topographic maps ^[54] and satellite images ^[55]. The same team carried out a very interesting experience in machine learning the recognition of ceramic fragments on a plowed surface for their subsequent mapping and spatial analysis, during which settlement structures are identified in places of concentration of lifting material ^{[56, 57]}. 

Another example concerns the modeling of the settlement of the territory of the Eastern Alps by Slavic tribes in the early Middle Ages ^[58]. A series of maps for narrow chronological periods was built by computer automatically based on the analysis of the database of archaeological sites of Central Europe using spatial GIS analysis by the method of "hot spots" (hot spot). The animation created by the authors conveys the gradual spread of settlements with a certain set of features associated with the cultural characteristics of the Slavic population.

The first examples of the use of artificial intelligence in archaeological research have also appeared in Russian science. Thus, V.M. Kostomarov and his co-authors managed to organize the training of a computer program for the recognition of untilled mounds on satellite images ^[59]. And a team of archaeologists and programmers from St. Petersburg is working on deep machine learning for automatic pattern recognition in order to facilitate the search for analogies to archaeological objects from the excavations of the IIMC RAS. For this task, the AIS "Archeographer" is used, which was discussed above ^[60]. Both of these examples were recently presented as reports at the Sixth International Conference "Archaeology and Geoinformatics" ^[52].

 Conclusion

The above examples, of course, are not limited to a wide range of areas of archaeological knowledge in which digital technologies are actively used. It is rather about some modern trends in computer archaeology that can be observed in recent years. Speaking about the absolute advantages of digital archaeology, which allows to accumulate and process huge amounts of information about archaeological landscapes, monuments, objects and finds, to create detailed and accurate three-dimensional models, to intensify routine processes of archaeological fixation, it is impossible not to mention some of the disadvantages and dangers that this area is fraught with. Usually, experts in the field of digital archaeology call the lack of a single developed methodological procedure in the field of digital fixation as one of the main weaknesses, as a result of which each team develops its own algorithm of actions for the use of digital technologies. Another dangerous moment, which is recognized by the scientific community, is associated with huge amounts of information that must be stored on an ongoing basis, which requires certain financial and organizational costs.

It seems to me that these problems are technically surmountable and solvable. The main danger that I see following some of the authors of the collection cited above ^[33] is that, being carried away by digital modeling, many researchers are gradually moving away from the main subject of their research – archaeological material that serves us to restore the historical picture of the past. Whatever digital models may be in their external attractiveness, they are only a tool in this process of cognition. A tool, of course, necessary and useful, but unable to completely replace other methods of archaeological science, such as, for example, stratigraphic or typological. David Clarke's "analytical machine" remains an impossible dream, because no computer, even with artificial self-learning intelligence, is yet able to completely replace the analytical abilities of archaeologists. Therefore, it seems useful to call some of our colleagues to the so-called "slow Archaeology" (Slow Archaeology) by analogy with "slow Food" (Slow Food), which is prepared individually and without haste for each individual consumer ^[61]. Perhaps this will be the further development of digital methods and their implementation in future archaeological research.