Home Geospatial Applications Archaeology Documenting mausoleums of the Chinese Tang Dynasty

Documenting mausoleums of the Chinese Tang Dynasty

Prof Dr Wolfgang Boehler


Prof Dr Wolfgang Boehler
Institute for Spatial Information and Surveying Technology,
University of Applied Sciences, Mainz, Germany
[email protected]

Dipl - Ing Guido Heinz


Dipl – Ing Guido Heinz
Roemisch-Germanisches Zentralmuseum, Mainz, Germany
[email protected]

The Tang project
The 18 mausoleums of the Chinese Tang Dynasty (618-907) are scattered over an area of 5.000 km² in the Shaanxi province in the PR China. The size of a single mausoleum, including the surrounding wall, may reach up to 15 km². Nowadays only mounds and single findings remain from the former buildings, walls and towers, whereas numerous larger-than-life stone sculptures are still in good condition. 

In 1993, cooperation between the Shaanxi Archaeological Institute, PR China, and the RGZM was established for the documentation of the mausoleums. i3mainz is a partner in this project and carries out all tasks concerning surveying and geoinformatics.


Fig. 1: Perspective view showing a part of a Tang Emperors’s mausoleum in Shaanxi, China. IKONOS Data have been draped over a DEM derived from a local tacheometric survey. Vector data from archaeologic prospecting (findings, tumuli, location of an ancient wall) are superimposed

Image data
During the course of working on the project, we had the oppurtunity to acquire and to evaluate many different images from the same area. Among those are:

  • LANDSAT TM
  • ERS-1 (Radar)
  • SPOT PAN 
  • SPOT PAN stereo
  • SPOT XS
  • IRS-1C PAN 
  • IRS-1C LISS III 
  • KVR-1000
  • IKONOS PAN
  • IKONOS multispectral 

Reference system
A common reference coordinate system, preferably the state coordinate system, must be used for all surveying and mapping activities. This allows the relocation of any site and a correct union with former and future documentation. A common system is also required when different data sources, such as maps, remote sensing, photogrammetry, and local field surveys, have to be combined. 

In the Tang project, a custom designed coordinate system is used for all ground surveys, as no information about the state coordinate system was available. The spatial positioning was determined using GPS. 

GPS receivers have also been successfully used in our project to supply coordinates for remote sites and control points. When brought together within a common reference system, data can be visualised and analysed in various combinations for use in research, publications, and museum presentations.

Table 1:Main specifications of sensors used
Sensor Spatial Resolution [m] Number of Channels Channels Comment
IKONOS PAN 1 1 Visible  
IKONOSMultispectral 4 4 B, G, R, NIR True colour possible
KVR-1000 2 at best 1 Visible Film
IRS-1C PAN 6 1 Visible  
SPOT PAN 10 1 Visible Orbit-Stereo
IRS-1C LISS III 18 3 G, R, NIR  
SPOT XS 20 3 G, R, NIR  
LANDSAT 32 7 B, G, R, NIR, IR, IR, TIR True colour possible
ERS 20 – 30 1 RADAR  

Stereo / DEM
A SPOT PAN stereo model was used to generate a digital elevation model of the target area. GPS was used for the determination of control points. The image data with a resolution of about 12 m allow to calculate an elevation model with about 50 m spacing and an accuracy of about 10 m. This model is suitable for the visualisation of extended mausoleums. The accuracy is not sufficient for the generation of maps in scales of 1 : 5,000 for the area of one mausoleum. Hence, it cannot replace the tacheometric measurements in the field to record topography.

Multispectral Data / Image Merging
Multispectral data can be used for landuse classification and natural coloured images. Landuse maps are suitable for overlays in views showing the current situation in combination with historical findings. The standard product for classifications is LANDSAT data with seven bands from visible to thermal infrared. 

Colour images are very useful for a natural visualisation of the site as they are more attractive as compared to black and white images. They can be produced if a red, green, and blue band (RGB) is supplied by the sensor. The combination with panchromatic data with a higher resolution improves the colour image. Several methods, including RGB-IHS-RGB transformation, principle components transformation and sub-pixel merging have been used and tested in this project for the combination of SPOT PAN with LANDSAT TM RGB as well as the panchromatic and RGB bands of IKONOS.  

Combination of Image Data with Ground Surveys
With the aid of image processing software, different types of data can be combined to show relations between information from different sources, e.g. archaeological objects, topography, and present landuse. Point, line, and area information from maps or local surveys can be visualised as vector symbols with CAD tools and combined with orthoimages or perspective views from satellite images (cf. Fig. 1). As mentioned above, it is essential to reference all information involved in a common coordinate system.

Using Satellite Image Data for planning and orientation in fieldwork
Raw digital or film images, as supplied by the distributors, can already be very helpful as a source of information about the project site and the surrounding area. The location of findings in relation to modern infrastructure like cities and roads can roughly be measured and used to plan further field work. Using simple plots of the image together with its approximate georeference as supplied by the distributor in combination with a simple handheld GPS receiver is very useful for orientation in the field and the determination of point positions in remote areas where maps are poor or not available. For this purpose, high accuracy is not essential, so the uncertainty in the geographic position of the image and in the GPS positioning is acceptable in most cases.

Mapping topography around the site
Any archaeological fact can only be explained and understood if the surrounding geographical facts are taken into account. In the case of the Tang mausoleums, geomancy played an important role for the selection of the sites and their arrangement. Nevertheless, the mapping of topography around archaeological sites is often neglected in documentation. An important reason for this is the comparatively large amount of work necessary. The use of remote sensing data can be a good compromise between expenditure and benefit, especially when other data, such as aerial images, are not available. In the current project, the use of standard data like SPOT and LANDSAT TM are sufficient for most mapping and visualisation tasks. Compared to standard mapping procedures, a lot of time and work in the field can be saved. 


Fig. 2: Huiling Mausoleum (ca. 30 m in diameter) in different images Upper row from left to right: IKONOS PAN, IKONOS MS, KVR-1000 Lower row from left to right: SPOT PAN, IRS-1C LISS III, LANDSAT

Mapping the Site / Replacing Ground Surveys
Obviously, detection of objects depends on pixel size. The reduction in pixel size has pushed the limit towards smaller objects. Single findings of small size and topographic detail plans still will have to be recorded locally in conjunction with the archaeological survey. In maps with smaller scales, a part of the field work can be replaced by remote sensing techniques. For these tasks the ground resolution of the data is a most important factor. Figure 2 shows Huiling Mausoleum with a mound of about 30 m in diameter and 15 m in height in different images. It is visible only in the high resolution images due to topography and the effects of light and shadow. Smaller objects, like stone sculptures with a base area of about 1.5 m x 3 m can only be noticed in the IKONOS PAN images, if their location is known. In the Tang project, the localisation of the numerous tomb mounds of officials and relatives consumes a lot of time. Many of these are not documented in Chinese literature and have to be searched for in the field. Since their diameters area at least around 5 m, satellite images can be used to locate these features by visual inspection or automated detection using image processing techniques.

Visualisation of archaeological objects
The most important point in using remote sensing image data for visualisation is the image information itself. A recent image of the region often brings much more information as compared to standard topographic maps. The combination with ground surveys and landuse information leads to a highly meaningful document. The combination with digital elevation models adds the topographic component and understanding. Almost photorealistic perspective views and animations can give an excellent impression of the region, the sites and findings, and the relations between them (cf. Fig. 1). The required image resolution is dependent on the scale of the representation (cf. Table 2). 

 Table 2:Comparison of images with different resolutions for different documentation tasks.
Task IRS-1C IKONOS
Planning and orientation Good for overviews and planning Not necessary
Mapping environs of sites For scales up to 1 : 20.000 For scales up to 1 : 5.000
Mapping sites For scales up to 1 : 10.000Objects of about 15 m size can be recognised. For scales up to 1 : 2.000objects of about 5 m size can be recognised.


Fig. 3: Query result of an archaeological GIS.

Information Systems for archaeological documentation
In the Tang project, a desktop GIS (ArcView) was used to examine the potential of such a system (Heinz, 1997). In the documentation of such extended sites enormous amounts of data have to be managed. This can sensibly be done using database and information systems. As the geographical and topographical location of all findings are an important attribute and topographic dependencies are of importance for the relation among the findings, the use of GIS tools can improve data storage and retrieval (Fig. 3). GIS allows the storage of different data types like vector drawings, tabular data, image data, text information, maps, elevation models, photos, etc.
By managing the documentation in a GIS, different data queries can be made possible:

  • Standard queries, e.g. listing of all lion sculptures looking westwards. 
  • Queries taking into account the location, e.g. max. distance to certain positions. 
  • Queries taking into account the topography, e.g. landscape slope in a certain direction, so geomantic aspects in planning the mausoleums can be evaluated. Visualisation of query results in existing maps. 
  • Identification of certain objects in the maps and displaying affiliated object properties. 
  • Access to further information about the objects like drawings, descriptions, photos, photogrammetric maps (Fig. 3). 
  • Additional representations like 3D-views or animated flights over the landscape. 
  • Combination of documentation results with present infrastructure by using satellite image data as background information.

Conclusions
Satellite image data are suitable for the documentation of extensive sites. Images with a standard resolution of about 10 m can be used for visualisations in map scales up to about 1 : 25.000. High resolution data (1 m) is suitable in scales up to 1 : 5.000. These images can match the quality of medium scale aerial photogrammetry. If multispectral data (usually with poorer spatial resolution) are available, they are useful in addition to panchromatic data for visualisations and animations in natural colour. The combination of image information and locally recorded single findings in maps helps to understand the topography of the site and can reduce the amount of time necessary for documentation on the site. The visibility of archaeological findings is obviously limited to objects of about pixel size. Ruins of buildings or long walls can be found in high resolution data whereas single sculptures are not visible. Since high resolution images from different systems are now available on the commercial market, large archaeological sites and their environs can be mapped, visualised and documented in GIS. This opportunity will not replace field work on the ground, but it can speed up the documentation process allowing the field personnel to concentrate on the investigation of the historical remains. 

References

  • www.i3mainz.fh-mainz.de 
  • Boehler, W., Heinz, G., Scherer, Y., 1997. Using Satellite Images for Archaeological Documentation. International Archives of Photogrammetry and Remote Sensing, Volume XXXII, Part 5C1B, pp. 226-233. 
  • Boehler, W., Heinz, G., 1999. Integration of High Resolution Satellite Images into Archaeological Documentation. International Archives of Photogrammetry and Remote Sensing, Volume XXXII, Part 5W11, pp. 166-172.
  • Boehler, W., Heinz, G., Scherer, Y., Siebold, M., 2001. Topographic Information in Cultural and Natural Heritage Visualization and Animation. International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIV, Part 5W1, pp. 56-61. 
  • Gong Qiming, Koch, A., 2002. Das Qiaoling. Gong Qiming: Bericht über archäologische Sondagen und Vermessungen am Qiaoling, Mausoleum des Tang-Kaisers Ruizong. Alexander Koch: Mausoleum des Tang-Kaisers Ruizong (662-716) im Kreis Pucheng (Prov. Shaanxi, VR China).
  • Heinz, G., 1997. Aufbau eines Geo-Informationssystems zur Dokumentation archäologischer Befunde. Tagungsband, 2. Geosystems Fachtagung, Germering.