多时相双极化合成孔径雷达干涉测量土地覆盖分类方法

摘要

In order to study SAR land cover classification method,this paper uses the multi-dimensional combination of temporal,polarization and InSAR data.The area covered by space borne data of ALOS PALSAR in Xunke County,Hei longj iang Province was chosen as test site.A land cover classification technique of SVM based on multi-temporal,multi-polarization and InSAR data had been proposed,using the sensitivity to land cover type of multi-temporal,multi-polarization SAR data and InSAR measurements, and combing time series characteristic of backscatter coefficient and correlation coefficient to identify ground objects.The results showed the problem of confusion between forest land and urban construction land can be nicely solved,using the correlation coefficient between HH and HV,and also combing the selected temporal,polarization and InSAR characteristics.The land cover classification result with higher accuracy is gotten using the classification algorithm proposed in this paper.%综合采用时相、极化和干涉3种维度的 SAR 数据进行土地覆盖分类。以黑龙江省逊克县多时相 ALOS PALSAR 数据覆盖区为研究区，利用不同时相极化 SAR、干涉 SAR 信号对地物特征的敏感性，结合后向散射强度和干涉相干的时变特征进行地物解译，发展了基于多时相、多极化、干涉 SAR 数据的SVM 土地覆盖分类方法。研究结果表明，引入双极化 SAR 中不同极化(HH-HV)间的相干系数，并结合所选择的时相特征、极化特征以及干涉相干特征进行分类，可解决双极化 SAR 影像中林地与城市及建设用地的混分问题，得到更高精度的土地覆盖分类结果。