Environmental Changes And Archaeological Study Using Multi-source Remote Sensing Datasets Of Typical Area In West China

China, as one of the oldest countries in the world, is the country with continuous historical sequence. There are many archaeological resources all over the country. The west China spans two drainage areas of Yellow River and the Yangtze River related to human civilization. However, because of extensive area, there are many topographical types such as canyon, plateau, basin, desert and so on, which make the natural environment complex and the climate terrible. Considering the above conditions, it is very difficult to investigate and study archaeological sites with traditional methods. Remote Sensing (RS) and Geographic Information System (GIS) provide a new method for archaeological research. Therefore, RS becomes an irreplaceable tools to explore large scale archaeological sites in west China.This research selected Tongwan City of Shanxi Province, the Ming Great Wall of Yulin and Sanxingdui of Sichuan Province as study areas, used multi-source remote sensing data (Landsat TM/ETM+, SPOT, Quickbird, IKONOS, etc.), related documents and geographical data, to locate and analyze the archaeological sites in study areas. The involved researches are as follows:Firstly, using Landsat TM data, we analyze the band range and spectral character of objects. The two methods-band ratio and factor index, are compared. Based on the result, we use factor index (SAVI, Cap transform, NDMI) method to extract environment information from the TM images of 1991 and ETM+ images of 1999. Using quantitative method, the environment changes within 10 years are analyzed and compared.Using high-resolution data ?Quickbird, we study the location, range and formation of Tongwan City and measure the length of city wall with GIS software-Arc View. According to the information acquired from remote sensing images and contour map, we analyze the reason of environment changes in Tongwan City.For the distributing character of Great Wall in desert, a new algorithm to extract Great Wall from remote sensing imagesis developed, based on the image gray level gradient. The result of experiment in IKONOS image is presented which show the effectiveness of the new algorithm.Because of the high resolution of IKONOS and Quickbird images, we use them to interpret and analyze the Ming Great Wall in Yulin City and remained wall in Sanxingdui. And we analyze the reason of environment deterioration near the Great Wall with NDVI.