LAND COVER CLASSIFICATION BASED ON TEMPORAL BACKSCATTER SIGNATURES OF THE TARGETS
SHAO Yun, FAN Xiang-tao, LIU Hao (Lab of Remote Sensing Information Sciences, Institute of Remote Sensing Applications, Chinese Academy of Sciences, Beijing 100101, China)
This paper presents the results of a study examining the backscatter signatures of various targets as a function of time. The characteristics of a target may show seasonal changes. If so, the backscatter signatures of the target should vary as a function of time. This serves as the basis of using multi-temporal SAR images to discriminate and classify a variety of targets. This study was carried out in Zhaoqing test site in Guangdong Province of China. The geometric characteristics of various targets and their backscattering mechanism were analyzed. The temporal backscatter of targets was emphasized and the backscatter signatures of targets as a function of time were summed up. Twelve types of land cover were classified using multi-temporal Radarsat data and, in addition, a land cover map was produced based on the classification results.
【CateGory Index】： TP79