COMPARATIVE ANALYSIS OF SEVERAL TYPICAL APPROACHES OF DESERTIFICATION DYNAMIC MONITORING BASED ON TM DATE

摘要

Land desertification is one of the most important problems about the issue of environment, social and economy in the world. Vegetation category and quantity is a significant criterion to measure the extent of land desertification, so desertification dynamic monitoring based on remote sensing depend on the distribution of vegetation. This paper take Huojitu mining area of the Erdos city, Inner Mongolia for example, extracting vegetation information by using methods of Normalized Difference Vegetation Index、 scaled difference vegetation index and interpreting mixed pixels, based on three temporal Landsat remote sensing images. Meanwhile, relative analysis of these three methods has been made. The results showed that: Vegetation Indices behaved the best in the region of high coverage. However, Pixel Unmixing Models is better suitable for dynamic monitoring vegetation in low coverage area. The SDVI, which is a linear index, was found to be a more suitable approach for retrieval of vegetation fraction. Besides, this paper systematic analysis Time-space evolution tendency of desertification land in Western China, also puts forward some preventive and controlling measures against sandy desertification.