Research on quantitative extraction of desertification integrating multi-source remote sensing data in large opencast coal mine area

摘要

Large opencast coal mine area is a complex region which consists of mining area, grassland, farmland, roads, residents and other landscapes, together with the typical characteristics of desertification, agro-pastoral transitional zone, mineral development area and other ecologically fragile areas. As a man-caused disturbance, mining plays an important role in desertification development in this area. The quantitative extraction of desertification integrating multi-source remote sensing data in large opencast coal mine area was studied in this paper. First, the remote sensing data from SPOT 5 and TM were fused to one image based on IHS transformation and Wavelet analysis, then the BP neural network algorithm were used to fusion image to get desertification classification information. The Global Positioning System (GPS) technique was also used to perform verification during the data process. The results of our studies indicate that the severe desertification is always distributing in shape of triangle around the mining land, which has much spatial correlation with the mining land. So the desertification management should be carried out in mining land, the neighbor region of the mining land should not be neglected.