文献类型：会议论文

英文题名：Classification method validation for rice mapping using envisat ASAR APS data

作者：Chen, Erxue[1] Li, Zengyuan[1] BingxiangTan[1] He, Wei[1] Li, Bingbai[2]

第一作者：Chen, Erxue

通信作者：Chen, E.

机构：[1] Institute of Forest Resources Information Techniques, Chinese Academy of Forestry, 100091, Beijing, China; [2] Institute of Agriculture Modernization, Jiangsu Academy of Agriculture Sciences, 210014, Nanjing, China

会议论文集：Proceedings of PolInSAR 2007: 3rd International Workshop on Science and Applications of SAR Polarimetry and Polarimetric Interferometry

会议日期：January 22, 2007 - January 26, 2007

会议地点：Frascati, Italy

语种：英文

外文关键词：Data reduction - Image acquisition - Maximum likelihood - Polarization - Satellites

年份：2007

摘要：ENVISAT is the first satellite that provides alternative polarization (AP) SAR data to end users. In order to validate the rice mapping capability of ASAR AP data, multi-temporal ASAR W-VH alternative polarization single look complex (APS) data products were acquired for the test site in Xinghua district of Jiangsu province. The three scenes of APS image were acquired in June 22 July 8 and Oct. 15, 2003. The Institute of agriculture modernization of Jiangsu academy of agriculture science was carrying an operational rice mapping project using Landsat TM data and other sources of ground truth data. The rice mapping results for the year 2003 was used as ground truth for this study. Two kinds of preprocessing methods were validated: backscattering coefficient based method (BSCBM) and the Polarimetric SAR data processing method (PSDPM). Different combinations of the output images from the two methods were used as inputs to a Maximum Likelihood Classifier (MLC). It has been observed that the performance of PSDPM is better than that of BSCBM; Multi-temporal W co-polarization SAR data has higher capability for land cover classification than multi-temporal VH cross-polarization SAR data; Integrating alpha and entropy images with all the 6 intensity images can achieve highest rice classification accuracy, but a litter bit lower total accuracy. Applying PSDPM to APS data and combining all the information from it as inputs to a certain classifier was suggested for operational rice mapping using multi-temporal ASAR APS data.