李梅,沈麒凯,陈启浩,刘修国

• 1:中国地质大学(武汉)地理与信息工程学院

摘要(Abstract)：

针对合成孔径雷达(SAR)图像不同类型建筑物的区分问题，该文提出了一种基于U-Net的极化SAR图像建筑物分类方法。该方法将极化SAR数据的Pauli分解参数、规范化圆极化相关系数和G~0统计纹理参数作为U-Net的输入，建立建筑物分类U-Net模型，同时考虑建筑物的高度和单体面积的情况下，将建筑物分为高层、中层、低层小面积、低层厂房类大面积建筑物4类。对武汉市城区GF-3极化SAR数据的各类建筑物分类精度均在80%以上，最高达94.2%。该方法与仅使用单类别特征的U-Net网络方法以及卷积神经网络方法相比，分类结果更完整、建筑边界更准确，也更适合于复杂中心城区的不同类型建筑物的分类。

关键词(KeyWords)： 极化SAR;U-Net网络;分类;建筑物;深度学习

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(41771467)

作者(Author): 李梅,沈麒凯,陈启浩,刘修国

DOI: 10.16251/j.cnki.1009-2307.2022.09.018

参考文献(References)：

• [1]吴文福，邵振峰，杨会巾．全极化SAR图像的建筑物提取研究[J]．测绘科学，2019,44(11):143-147.(WUWenfu,SHAO Zhenfeng,YANG Huijin.Extraction of buildings from full-polarimetric SAR imagery[J].Science of Surveying and Mapping,2019,44(11):143-147．）
• [2]王贺，张路，徐金燕，等．面向城市地物分类的L波段SAR影像极化特征提取与分析[J]．武汉大学学报（信息科学版），2012,37(9):1068-1072.(WANG He,ZHANG Lu,XU Jinyan,et al.Evaluation of L-band polarimetric SAR images for urban land cover classification[J].Geomatics and Information Science of Wuhan University,2012,37(9):1068-1072．）
• [3]陈启浩，聂宇靓，李林林，等．极化分解后多纹理特征的建筑物损毁评估[J]．遥感学报，2017,21(6):955-965.(CHEN Qihao,NIE Yuliang,LI Linlin,et al.Buildings damage assessment using texture features of polarization decomposition components[J].Journal of Remote Sensing,2017,21(6):955-965．）
• [4]刘修国，姜萍，陈启浩，等．利用改进三分量分解与Wishart分类的极化SAR图像建筑提取方法[J]．测绘学报，2015(2):206-213.(LIU Xiuguo,JIANG Ping,CHEN Qihao,et al.Buildings extraction from polarmetric SAR image using improved three-component decomposition and Wishart classification[J].Acta Geodaetica et Cartographica Sinica,2015,44(2):206-213．）
• [5]MORIYAMA T,YAMAGUCHI Y,URATSUKA S,et al.A study on polarimetric correlation coefficient for feature extraction of polarimetric SAR data[J].IEICETransactions on Communications,2005,88(6):2353-2361．
• [6]AINSWORTH T L,SCHULER D L,LEE J S.Polarimetric SAR characterization of man-made structures in urban areas using normalized circular-pol correlation coefficients[J].Remote Sensing of Environment,2008,112(6):2876-2885．
• [7]ZHAI Wei,SHEN Huanfeng,HUANG Chunlin,et al.Fusion of polarimetric and texture information for urban building extraction from fully polarimetric SARimagery[J].Remote Sensing Letters,2016,7(1):31-40．
• [8]FREITAS C C,FRERY A C,CORREIA A H.The polarimetricgdistribution for SAR data analysis[J].Environmetrics,2005,16(1):13-31．
• [9]LI Linlin,LIU Xiuguo,CHEN Qihao,et al.Building damage assessment from PolSAR data using texture parameters of statistical model[J].Computers&Geosciences,2018,113:115-126．
• [10]AZMEDROUB B,OUARZEDDINE M,SOUISSI B.Extraction of urban areas from polarimetric SAR imagery[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2016,9(6):2583-2591．
• [11]姜萍，刘修国，陈启浩，等．利用多尺度SVM-CRF模型的极化SAR图像建筑物提取[J]．遥感技术与应用，2017,32(3):475-482.(JIANG Ping,LIU Xiuguo,CHEN Qihao,et al.A multi-scale SVM-CRF model for buildings extraction from polarimetric SAR images[J].Remote Sensing Technology and Application,2017,32(3):475-482．）
• [12]DU P J,SAMAT A,WASKE B,et al.Random forest and rotation forest for fully polarized SAR image classification using polarimetric and spatial features[J].ISPRS Journal of Photogrammetry and Remote Sensing,2015,105:38-53．
• [13]DE S,BRUZZONE L,BHATTACHARYA A,et al.Anovel technique based on deep learning and a synthetic target database for classification of urban areas in PolSAR data[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2017,11(1):154-170．
• [14]ZHOU Yu,WANG Haipeng,XU Feng,et al.Polarimetric SAR image classification using deep convolutional neural networks[J].IEEE Geoscience and Remote Sensing Letters,2016,13(12):1935-1939．
• [15]ZHANG Zhimian,WANG Haipeng,XU Feng,et al.Complex-valued convolutional neural network and its application in polarimetric SAR image classification[J].IEEE Transactions on Geoscience and Remote Sensing,2017,55(12):7177-7188．
• [16]ZHANG Ce,PAN Xin,LI Huapeng,et al.A hybrid MLP-CNN classifier for very fine resolution remotely sensed image classification[J].ISPRS Journal of Photogrammetry and Remote Sensing,2018,140:133-144．
• [17]WU Wenjin,LI Hailei,LI Xinwu,et al.PolSAR image semantic segmentation based on deep transfer learning:realizing smooth classification with small training sets[J].IEEE Geoscience and Remote Sensing Letters,2019,16(6):977-981.