基于高分三号全极化数据的西宁市城区地表地物变化检测  

Detection of surface feature changes in Xining city based on GF-3 full polarized data

作　　者：靳生洪 辛兵厂[1] 史磊[2] 刘元旭 JIN Shenghong;XIN Bingchang;SHI Lei;LIU Yuanxu(Provincial Geomatics Center of Qinghai,Xining 810000,China;State Key Laboratory of Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan University,Wuhan 430000,China;Geomatics Technology and Application key Laboratory of Qinghai Province,Xining 810000,China)

机构地区：[1]青海省基础地理信息中心,青海西宁810000 [2]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430000 [3]青海省地理空间信息技术与应用重点实验室,青海西宁810000

出　　处：《青海大学学报（自然科学版）》2019年第1期79-86,共8页Journal of Qinghai University(Natural Science)

基　　金：青海省地理空间信息技术与应用重点实验室开放基金资助项目(QDXS-2017-04)

摘　　要：为了深入研究国产高分三号(GF-3)数据在测绘领域的适用性,科学评价数据的应用价值,文中基于青海省西宁市2017年9月21日和10月20日双时相全极化数据,在图像配准的基础上,采用H-Alpha-Wishart非监督分类法对西宁市城区地表地物进行了分类与变化检测研究,评价了GF-3数据在地物分类与变化检测中的可用性。试验证明:GF-3的全极化数据质量较为稳定,且GF-3分类精度优于90%、Kappa系数0.86,可以用于城区、植被、布拉格(Bragg)散射目标等地物的分类与变化检测,对自然资源调查以及常态化监测等相关应用具有一定实用价值。In order to deeply study the applicability of domestic GF-3 data in surveying and mapping field,the application value of GF-3 data is evaluated scientifically,this paper based on the dual-phase full polarization data of September 21 and October 20,2017 in Xining city,Qinghai province,the classification and change detection of surface features in Xining city was studied by using H-Alpha-Wishart unsupervised classification method on the basis of image registration,and the availability of GF-3 data in surface features classification and change detection was evaluated.Our experiments demonstrate that the quality of GF-3 data is stable and the classification accuracy of GF-3 is better than 90% and Kappa coefficient is 0.86.It could be considered as reliable data source to distinguish city areas,vegetation and bragg targets for classification and change detection,which has certain practical value to natural resources surveying and normalized monitoring and other related applications.

关 键 词：高分三号 合成孔径雷达 变化检测 西宁市 

分 类 号：P231.5[天文地球—摄影测量与遥感] TN957.52[天文地球—测绘科学与技术]