干旱区生态环境知识资源中心

Vegetation patches are one of the important components of arid and semi-arid ecosystems. Monitoring and mapping vegetation patches are essential for studying its succession mechanism. This study is aimed at comparing the differences between SPOT-5, ALOS and ZY-3 imagery and then showing the potential of them in recognizing vegetation patches. In this study, we deal with the three kinds of images with the method of object-oriented classification so as to extract circular or elliptical vegetation community patches of the two study areas. The results show that SPOT-5 imagery is of the highest segmentation accuracy, followed by ZY-3, and then ALOS. Then detailed analysis is carried out from four perspectives, which are statistical results of bands, NDVI distribution, image information entropy and sharpness and the conclusion is that the segmentation accuracy of each image is related to its NDVI distribution and articulation. Besides, ZY-3 imagery has the most balanced wave energy, the highest grey level and the richest information of the three. With good display effect and low price, ZY-3 imagery will be of great application potential in the field of remote sensing of vegetation.