UDC 902.26

The paper is devoted to studying the experience of using GIS technologies and remote sensing data together to create archaeological maps. In the course of the research, the main task was solved: the technology of digital mapping of archaeological sites was developed and tested. Digital map layers have been prepared for the Yustyd, Ulandryk I and II, and Sary-Gabo burial grounds. Digital terrain models have been created for them, which are used to analyze settlement systems, territorial organization of the economy, and similar reconstructions. A digital cartographic framework has been prepared, including a cartographic database and identification characteristics of monuments. The possibility of using DD for solving problems of archeology is shown. Algorithms for automated decoding of multispectral images for recognition, recording, and description of archaeological sites and their attributes have been improved and refined. The obtained materials and survey results (topographic and space scanner surveys) are presented in the most common GIS formats MapInfo Pro and ArcView.

Keywords: digital mapping of archaeological sites, GIS technologies, remote sensing in archeology, scanning, multispectral image decoding.

Introduction

One of the most acute problems of Siberian archeology is the lack of a detailed archaeological map of Altai. For its development, issues related to the development of the concept of geoinformation mapping in archeology are of great importance; the creation of a unified methodology that allows obtaining archaeological maps of the required scale and territorial coverage; the introduction of common standards and formats for digital cartographic data in archaeological research, etc. With the current level of development of geoinformatics and remote sensing methods, it is possible to quickly build archaeological maps with high accuracy, detail and reliability. This paper examines the experience of using GIS technologies and remote sensing data together to create an archaeological map for the Chui basin site.

In territorial terms, the object of our research is the South-Eastern Altai - the valleys of the Yustyt, Ulandryk, and Barburgazy rivers (Fig. 1). The region is characterized by peculiar mountain-steppe landscapes. This territory is not without reason considered an archaeological reserve of our country. Many cultural and historical monuments of different eras are located here. Among them, the most famous were the mounds of the Scythian period. However, until the 1960s, the Southeastern Altai remained a blank spot for archaeologists: not a single mound was excavated (Kubarev, 1987, 1991). Since the 1970s, studies of a large number of ordinary burial mounds in the Chui Steppe have been actively conducted [Kubarev, 1987, 1991, 1992]. But until now, the South-Eastern Altai, and in particular the territory of the Chui basin, have not been sufficiently studied, since surveys were conducted within separate territorial areas and burial grounds.

This work was supported by the Russian State Scientific Foundation, project No. 04-01-00470a " Ancient nomads of Altai: structure, functioning and development of settlement systems (I millennium BC)".

1. Ulandryk River valley. Photo by E. P. Krupochkin.

Fig. 2. Olenny kamen. Right bank of the Yustyt river. Photo by I. Y. Slyusarenko.

Moreover, most of the burial grounds and especially isolated archaeological sites (kereksurs, ritual fences, stone mounds, etc.) still do not have a clear geographical and coordinate reference.

The presented work is the result of joint research of the Dendrochronological Group of the Institute of Archeology and Ethnography of the Siberian Branch of the Russian Academy of Sciences (Novosibirsk) and the Faculty of Geography of the Altai State University (Barnaul). The main goal of the research is to develop a technology for digital mapping of archaeological sites on the example of the Chui basin. The mapping process involves recording archaeological sites, drawing up digital topographic plans of burial grounds, preparing and correcting digital relief models, and geodesic verification of the orientation of mounds on previously studied burial grounds (Ulandryk, Yustyd, Barburgazy, etc.) to analyze the orientation properties of monuments. The relevance of the topic is determined, on the one hand, by the lack of a well-developed technology for digital mapping of archaeological sites, and, on the other hand, by the development and need to introduce new methods of geoinformatics and remote sensing in the field of archeology.

Materials and methods

Let's consider the methodological aspects of mapping monuments on the example of the archaeological microdistrict "Yustyd". Archaeological research in the Yustyt River valley began in the late 60s of the XX century. The most significant of them were carried out by the Altai detachment under the leadership of N. M. Zinyakov and the East Altai detachment of the North Asian Expedition of the Institute of Scientific Research of the Siberian Branch of the USSR Academy of Sciences under the leadership of V. D. Kubarev (Bykova, 2002). During the period from the late 1960s to the early 1980s, entire burial grounds and individual mounds (Yustyd I-XXII) were discovered and studied, and iron-making monuments, pottery workshops, and deer stones were recorded on the right bank of the Yustyd River (Fig. 2). However, the problem of mapping cultural and historical monuments of Yustyd remains unknown.the problem remains unresolved.

In 2004, a Belgian group of researchers from the University of Ghent proposed a solution to this problem [Goossens et al., 2006]. The survey was carried out along the right bank of the Yustyt River from the Kalan spring to the main road leading to the village. 3). In the course of the work, the following tasks were solved:: 1) develop and test a methodology for mapping archaeological sites based on the use of remote sensing data of the CORONA satellite system; 2) record and plot on a single map all marked objects of the Bronze, Iron and Medieval eras; 3) supplement the UNESCO World Heritage List with objects of archeology of the South-Eastern Altai.

From the point of view of archaeological research, the main task in remote sensing is image recognition, i.e. the decryption process involving the use of special software packages (such as Photomod, ENVI, etc.) and geographic information systems that support working with bitmap images (Maplnfo Pro, ArcGIS, etc.) [Krupochkin, 2004]. As a rule, the original negatives of images have a number of distortions caused by various factors. -

3. Map coverage scheme of the archaeological microdistrict "Yustyd".

a - GPS survey area of the Belgian group of researchers (2004); b-GPS points; c-research area of the Dendrochronological Group of the Institute of Electric Power Engineering of the Siberian Branch of the Russian Academy of Sciences (2005-2008); d - field road.

Therefore, satellite images cannot be directly used in GIS without preliminary photogrammetric processing and transformation (orthophototransformation, calibration, binding to a geodetic coordinate system, etc.). The result of these works is images prepared in an orthogonal projection (without planned distortions), which can be used as a reference to the satellite's trajectory. in large-scale archaeological research.

One of the main operations for transforming an image is to snap it to a coordinate position. To clarify the spatial coordinates in the field, a network of GPS control points was formed (see Fig. 3) and selected geographically stable points in time. Three systems were used for reading coordinates: C-Nav (a high-precision geodetic system), a Garmin 12-channel pocket receiver (5 - 10 m positioning accuracy), and Motorola Oncore VP (15 - 30 m positioning accuracy). A greater effect was achieved by the first system, which is much more expensive. The other two can be used in some cases to process images only of objects with dimensions of 2 m or more.

Belgian studies have shown that the best results can be obtained by using satellite images from the Quickbird and Ikonos satellites (with a resolution of 0.6 and 1.0 m) (Figure 4). However, it is desirable to use an autonomous differential GPS system of the C-Nav type (Goossens et al, 2006).

4. Topographic plan of the Yustyd XIII burial ground.

A fragment from the general plan of Belgian researchers [Goossens et al., 2006].

The problem of creating digital archaeological maps of monuments was solved by us using a more economical technology. The initial stage of research included a reconnaissance, which was accompanied by a comprehensive archaeological and physical-geographical description of the area (the left and right banks of the Yustyt River Valley section) directly in the

5. Topographic plan of the Yustyd XIII burial ground based on field surveys carried out in 2005-2008 by the Dendrochronological Unit of the Institute of Electric Power Engineering of the Siberian Branch of the Russian Academy of Sciences.

- a non-excavated mound; b-excavated mound; c-memorial layout; d - control point of the planned high-altitude survey network; e - Turkic fence.

field. The coordinates of objects were recorded using two GPS navigators GarminEtrex (±5 m) and Garmin Mapx60 (±2 m). Factors affecting the accuracy of satellite observations were taken into account: mechanical obstacles, reflecting objects, radio interference, refractive effect, etc.

The next stage is total and theodolite surveys of burial grounds. The first differs from the second primarily in the speed of work performed, which does not require laying a theodolite passage. The survey was carried out using a high-precision theodolite 2T5K. At the same time, an outline of the terrain was created, on which the objects being shot with relief elements were recorded. Unfortunately, the territory under study is not covered by the State Geodetic Network (GGS). In addition, GGS stations created according to the Soviet GUGK standards, and even more so industry-specific ones, gradually "accumulate" the total displacement of the absolute coordinates of remote points up to many tens of meters (Postnov and Vergunov, 2003). In addition, free movement in the territory was restricted by two factors: first, it is a border zone, and secondly, it is represented by very complex terrain areas. Therefore, the geographical reference of the captured sites was performed at three points (the station and two additional extreme points located diagonally along the site), the coordinates of which were recorded using a GPS receiver. In the future, in the process of camera processing of field survey data, as well as for drawing up digital topographic terrain plans, software packages and modules were used: Credo DAT, GIS Maplnfo Pro, GIS application "Spatial Analyst" for Arc View (Fig. 5). To transfer data from GPS and ensure compatibility of GIS formats, all the presented solutions were used. cartographic materials were prepared in the universal geocentric system WGS-84 (developed in 1984 and currently used in the Navstar radio navigation system). For the Altai Republic, most topographic maps are made in the cross-cylindrical Gauss-Kruger projection SK-42. Among the cartographic projections that are most often used in GIS for a number of basic quasigeoid parameters, the WGS-84 projection is the closest to SC-42, which justifies its choice.

Along with drawing up topographic plans and archaeological maps, digital relief models (DEM)were prepared and corrected* burial grounds and other model sites (Krupochkin, 2007). The most common ways to organize and represent DEMS are raster data models and a special spatial data model based on the triangulation network (TIN), which approximates the terrain with a polyhedral surface with height markings at the nodes of a triangular network. Methods and algorithms for creating and processing DEMS are also applicable to other physical or statistical reliefs and fields: buried relief (in archeology), baric relief (in climatology), etc.

In modern geoinformatics and cartography, a distinction is made between digital elevation models (Dem-1) and their derivatives - terrain models (Dtm). In this case, Dtm refers to a set of morphometric indicators derived from the terrain. The need for a distinction is partly related to the name and content of the American DEM Standard (Dem-2) (Baranov et al., 1999).

* DEM - a means of digital mathematical representation of three-dimensional spatial objects (surfaces, reliefs) in the form of three-dimensional data as a set of elevation marks and other values of appendices (Z coordinates) in the nodes of a regular network with the formation of a height matrix, an irregular triangular network (TIN), or as a set of records of horizontals or other isolines [Baranov et al., 1999].

When creating digital terrain models, we used two data sources: elevation markers recorded in GPS Waypoints and calculated values of absolute picket heights obtained during camera processing of total station survey results. To solve the technical problem of DEM modeling and visualization, a functional approach was used, which assumes that the relief is described by a certain function approximating with the help of another function: where V. - angles of inclination of the surface, V₁Z_min-functional. The approximation allows us to study the numerical characteristics, qualitative (morphological) and quantitative (morphometric) properties of the terrain of the territory. To solve this problem, the condition = min must be met, where [f(x_ky_k) = Z] is a differentiating function. Since the result should be represented as a map, the coordinates in the plan (X, Y) were replaced with geographical coordinates (φ φ). Conversion to the geographical coordinate system was performed by projecting the axes of the coordinate system, where the Greenwich meridian is represented as χ=X, and the equator line as φ = Y.

Geodetic verification of the orientation of the mounds was carried out using previously studied burial grounds in the South-Eastern Altai (Ulandryk, Yustyd, Barburgazy, etc.). Its essence is to determine the geographical azimuths of the axes of log cabins, taking into account the measured magnetic azimuth and magnetic declination determined from a topographic map. To determine the seasonality of burials, an important information is the closed horizon. The higher this indicator is, the greater the geographical azimuth of sunrise. Measurements of the closed horizon were made for each mound using a theodolite with an accuracy of up to a minute. The dates of sunrise and sunset for a particular year were determined using the Redschift program according to the method proposed by N. I. Bykov (Bykov and Bykova, 2003; Bykov et al., 2004).

A comparative analysis of the two approaches, represented in the first case by extensive use of the Navstar radio navigation system and high-precision GPS stations, and in the second case by combining traditional survey methods with GIS technologies and remote sensing data, allowed us to draw certain conclusions.

1. It is necessary to evaluate the possibility of replacing (full or partial) the field method of observation and survey of archaeological sites with remote. This requires an estimate of the price ratio for ordering a new space survey (or receiving archived materials). on the territory of interest and the total estimated cost of research.

2. The technical parameters of the survey equipment (sensor) and the characteristics of the proposed satellite images should be taken into account. First of all, you need to pay attention to the spatial and radiometric resolution of the image, the level of photogrammetric processing, which involves performing the following procedures: orthotransformation (elimination of distortions in the image caused by terrain dissection), geometric and radiometric calibration (elimination of the influence of light differences due to the shooting geometry; elimination of image defects; computer correction of image brightness values; calculation of calibration coefficients, etc. etc.), converting the photogrammetric coordinate system of the image into a geodesic one, etc.

3. A detailed study of the current conditions for conducting space surveys and the observed trends in reducing the cost of remote sensing materials allow us to recommend a wider use of satellite images to ensure mass archaeological research of cultural and historical monuments.

The preference of using the methods of remote sensing and mathematical-cartographic modeling for solving problems of archeology is determined by a wide range of possibilities:

- obtaining cartographic material with a sparse load (compared to topographic maps of a comparable scale) ;

- ordering images with significant spatial coverage;

- obtaining and analyzing a small area digital image (up to 25 ^km2);

- mapping of archaeological sites that are usually not shown on topographic or other special maps and plans, or are shown with insufficient accuracy and detail;

- objectification based on mathematical and cartographic models (ancient population density fields, dynamics maps and "retrospective forecast") of reconstruction of the course of historical processes;

- studying the stages and forms of ancient human settlement, as well as dating burials using astronomical and geodetic information and phytoindication method;

- creation of digital archaeological plans and thematic maps;

- search and identification of new archaeological sites using automated decryption algorithms;

- spatial 3D and animation modeling of archaeological complexes in their interrelation with the natural and anthropogenic environment, etc.

Figure 6. Fragment of the synthesized image obtained as a result of processing the original satellite image using the "Gauss-Laplace" algorithm.

Certain results were obtained by comparing field observations and decoding data from the Quickbird multispectral image (a section of the Yustyt River valley). Let's look at the technological principles and main stages of decryption. Based on the available information in the image, we determined a priority method of decryption based on the use of not only RGB channels (R-red, G - green, B - blue), but also infrared (IR). This is due to: the absence of a panchromatic image with a high or relatively high (2.01 m) resolution at the moment; the peculiarities of "Mongolian-type" landscapes, where morphostructural elements of monuments can be clearly distinguished (rings, contours of fences, structural elements of kereksurs, platforms and fences of free-standing mounds, elevations and depressions, etc.); the lack of data stereo surveys that would significantly enhance the search and reveal the possibilities of typologizing archaeological sites based on three-dimensional mathematical and cartographic modeling of the relief of the studied territory.

The classification is based on the image obtained in the near-infrared spectrum, since the image we have is characterized by a narrow optical range: from 0.48 to 0.83 microns. Conversion of the original image from the IR range was performed using the general vegetation index (NDVI). At the first stage, the class conversion of the IR channel with respect to Red (red) was performed using the formula NDVI = IR + Red / IR - Red.

In the image obtained as a result of processing, light areas correspond to a higher index value, which means a more active vegetation period. Please note that when calculating vegetation indices for vegetation analysis, we do not get absolute values, but a relative assessment of a certain property of vegetation. The processed image shows only territorial differences, i.e. where vegetation activity is higher and where it is lower.

At the second stage, algorithms for controlled classification of the resulting image were used. Using a comparative analysis of the decryption results, the most effective algorithms were identified that allow us to distinguish characteristic local areas against the background of the general mosaic, showing the location of both individual mounds and their aggregates in the form of chains along the prevailing axes of directions (north-south, south-west-north-east). The principle of operation of the algorithms is based on the use of mathematical operators "Gauss-Laplace" and "Variable Filter", which are designed to smooth the image (remove "noise") and emphasize the boundaries of objects.

The image processed using the Gauss-Laplace algorithm (Figure 6) clearly shows the contours of monuments. According to direct and indirect decoding signs, it is clear that point 7 is a mound with a stone embankment. The embankment borders are blurred. The slopes of the mound are blackened, but dark stripes indicate the presence of areas with loose stones or homogeneous soil. The dark color in the center of the mound indicates the absence of vegetation, which is confirmed by field observations (the presence of several microbes with an average depth of 30 cm). In the southwestern part, the color of the predominant class of the spectrum does not stand out from the general background. From this, we can assume that the external contour of the item is broken here. Field observations have shown that there is indeed a spread of material in this part of the mound (Figure 7).

The Variable Filter algorithm also made it possible to select several objects in a multispectral image. For example, the chain of Pazyryk mounds with the direction of the south-east-north-west axis is clearly distinguished (Fig. 8, points 20-21). However, it is almost impossible to determine the presence of individual archaeological sites due to poorly expressed unmasking signs, as well as due to the small size of the mounds themselves in the chain (from 2 to 5 m in diameter). Other image classification methods that remove "noise" and improve image quality did not show positive results. Therefore, for similar search conditions and coordinate fixation of archaeological sites, we can recommend the "Variable Filter" algorithm, since it allows you to find

and display on the multispectral image structural elements expressed in morphological terms, but "hidden" for normal visual decoding.

An interesting result was shown by the algorithm for constructing a hyperspectral cube - "3D Cube". The three-dimensional function "3D Cube" allowed performing multispectral analysis with the creation of a synthesized file that can be analyzed both spatially and spectrally (meaning calculating statistical parameters of the brightness values of pixels of the obtained image matrices). As a result of applying this function, a combined image with a three - dimensional effect is obtained, then by connecting RGB channels-a pseudo-hyperspectral three-dimensional cube. The use of the "Build 3D Cube" algorithm showed significant efficiency of the analysis in relation to a binary (black-and-white) image, as well as to an image in the RGB range from 0 to 255 (color image or black-and-white with 256 grayscale options).

The resulting image has improved the unmasking features of archaeological sites (shape, size, tone, etc.) (Fig. 9); with their help, you can easily identify herexes located in open areas along field roads. The dimensions of mounds, planimetry (location in the plan), and morphology (external surface structure) are determined. For example, on some sites there is a significant excess in the center of Kereksur, expressed in relief, on others-significant depressions (depressions), which may mean a failure of the internal structure of the burial or the transfer of material. Objects 16 and 17 clearly show holes in the centers. On the periphery, an outer belt is distinguished, presumably composed of large stones that form a distinguishable contour.

Further processing of the image involves applying already known and developing new automated decryption methods based on the principles and approaches of controlled classification, i.e. creating training algorithms that the program uses to automatically search for and select classes "by sample". To increase the efficiency of digital mapping of monuments, it is necessary to develop and improve methods of mathematical and cartographic modeling. Such tasks can be solved by creating DEMS of removable sites.

7. The mound corresponding to the coordinates of point 7 of the satellite image. Photo by E. P. Krupochkin.

Figure 8. Fragment of the synthesized image obtained as a result of processing the original satellite image using the "Variable Filter" algorithm.

Figure 9. Fragment of the synthesized image obtained as a result of processing the original satellite image using the "Build 3D Cube" function.

Results and conclusions

In the course of the research, digital cartographic materials on the archeology of the South-Eastern Altai (Chui basin) were obtained. In particular, digital map layers have been prepared for the Yustyd, Ulandryk I and II, and Sary-Gabo burial grounds. Digital terrain models have been created for them, which can be useful for analyzing settlement systems, territorial organization of farms, and similar reconstructions.

The work was carried out in parallel with the preparation of topographic plans. In the process of creating electronic maps, the system of symbols, methods of cartographic representation of archaeological sites, and algorithms for automated decoding of multispectral images were developed and improved.

A complex of works on geoinformation mapping of archaeological sites was carried out on the example of the archaeological microdistrict "Yustyd". A digital cartographic framework has been prepared, including a cartographic database and identification characteristics of monuments. The initial materials for forming the Yustyd GIS were the results of GPS mapping and instrumental survey of monuments in 2005-2007, as well as the results of decoding the Quickbird multispectral satellite image of the upper part of the river valley. The data obtained is presented in the most common GIS formats Maplnfo Pro and ArcView.

Geodetic verification of the orientation of mounds on previously studied burial grounds in the Ulandryk and Yustyt valleys was performed to analyze the orientation properties of monuments. In the field, theodolite was used to measure the angles of horizon closure, which is important for determining the burial season.

The expedition work carried out in 2005 - 2008 in the Yustyta Valley made it possible to:

- identify and describe the archaeological and physical-geographical features of monuments by key sites;

- compare the obtained preliminary results of decoding the existing satellite image with the expected ones;

- evaluate the relative error of the applied algorithms and determine the further direction of scientific research to improve the methods and algorithms of decryption.

Taking into account the still rather significant cost of high-resolution satellite images, it is advisable to use them in a comprehensive manner, i.e. not only for determining the location and size

It can also be used to create digital terrain models, analyze landscape structures, and so on, which will increase their economic efficiency.

List of literature

Baranov Yu. B., Berlyant A.M., Kapralov E. G., Koshkarev A.V., Serapinas B. B., Filippov Yu. A. Geoinformatics: Explanatory dictionary of basic terms, Moscow: GIS-Association, 1999. http://www.gisa.ru/geoinfoslovar.html (accessed: 15.01.2008).

Bykov N. I., Bykova V. A. Orientation of buried people in the mounds of the Scythian epoch as a source of historical and geographical information. - 2003. - Vol. 6. - p. 214-226.

Bykov N. I., Bykova V. A., Panyushkina I. P., Slyusarenko I. Yu. Dendrochronological and geodesic-astronomical assessment of the sequence of construction of mounds in the burial grounds of the Pazyryk culture of Altai / / Complex studies of ancient and traditional societies in Eurasia. Barnaul: Alt. State University Publ., 2004, pp. 258-264.

Bykova V. A. History of the archaeological study of the Chuya basin // Geography and nature management of Siberia. - 2002. - Issue 5. - p. 245-255.

Kitov A.D. Computer analysis and synthesis of geoimages. Novosibirsk: Publishing House of the Siberian Branch of the Russian Academy of Sciences, 2000, 220 p.

Krupochkin E. P. Geoinformatsionnyi podkhod k kartografirovaniyu arkheologicheskikh pamyatnikov (na primere territorii Respubliki Altay) [Geoinformation approach to mapping archaeological sites (on the example of the territory of the Altai Republic)]. Novosibirsk: Publishing House of IAET SB RAS, 2004, vol. 10, part 2, pp. 218-222.

Krupochkin E. P. Opyt ispol'zovaniya GIS-tekhnologii dlya resheniya zadach kartografirovaniya arkheologicheskikh pamyatnikov [Experience of using GIS technologies to solve problems of mapping archaeological sites]. Severnaya Bvraziya v antropogene: chelovek, paleotechnologii, geoekologiya, etnologiya i antropologiya: Mat-ly Vseros.konf. from internad. participation dedicated to the 100th anniversary of the birth of M. M. Gerasimov. Irkutsk: Ottisk Publ., 2007, vol. 1, pp. 312-321.

Kubarev V. D. Kurgany Ulandryka [Mounds of Ulandryk]. Novosibirsk: Nauka Publ., 1987, 302 p. (in Russian)

Kubarev V. D. Kurgany Yustyda [Mounds of Yustyda]. Novosibirsk: Nauka Publ., 1991, 190 p. (in Russian)

Kubarev V. D. Saylyugem mounds. Novosibirsk: Nauka Publ., 1992, -20 p. (in Russian)

Postnov A.V., Vergunov E. G. Fundamentals of geodetic support for archaeological research using satellite navigation receivers. Novosibirsk: Svet Publ., 2003, 160 p. (in Russian)

Goossens R., De Wulf A., Bourgeois J., Gheyle W., Willems T. Satellite imagery and archaeology: the example of CORONA in the Altai Mountains // J. of Archaeological Science. - 2006. - N 33. - P. 745 - 755.

The article was submitted to the Editorial Board on 17.10.08.

Permanent link to this publication: