基于高分二号卫星影像的城市不透水面提取  

Urban Impervious Surface Extraction Based on the GF-2 Satellite Imagery

作　　者：刘波 张源[2] 程涛[2] 宋杨[3] LIU Bo;ZHANG Yuan;CHENG Tao;SONG Yang(Geomatics Center of Guangxi, Nanning 530023, China;State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430072, China;Guangzhou Urban Planning Survey and Design Institute, Guangzhou 510060, China)

机构地区：[1]广西壮族自治区基础地理信息中心,广西南宁530023 [2]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430072 [3]广州市城市规划勘测设计研究院,广东广州510060

出　　处：《地理信息世界》2017年第2期103-107,共5页Geomatics World

基　　金：2016年广州市科技计划项目(201604020070);地理国情监测国家测绘地理信息局重点实验室基金(2015NGCM);武汉市晨光计划人才项目(2016070204010114)资助

摘　　要：受到分类目标趋于多样化和影像因素复杂的影响,基于高分辨率遥感影像提取城市不透水面的方法普遍精度不高。本文旨在探索快速有效地利用高分辨率遥感影像提取不透水面方法。以梧州市高分二号遥感卫星影像为例,采用支持向量机方法应用于不透水面分类中。这种方法首先根据高斯核函数训练出样本空间,然后直接对经过HSL色彩空间变换后的影像进行分类,使得有效特征信息增加进而分类精度提高。实验结果证明了这种方法的有效性。The accuracy of impervious surface is not high because of strong heterogeneity and complicated factors.Therefore this paper aims to explore effective method to achieve high-precision urban impermeable surface extraction.In this paper,based on SVM model and HSL transform,the GF-2Satellite imagery was taken as the data source to complete the impervious surface extraction of Wuzhou City.Firstly according to the Gaussian Kernel function,sample space was created.Then image which has been processed with HSL transformation information was classified.Finally it was found that after HSL transformation,the overall accuracy were improved.The results were effective for thematic data for the planning and construction of sponge city.

关 键 词：不透水面 支持向量机 高分辨率遥感 HSL变换 

分 类 号：TP227[自动化与计算机技术—检测技术与自动化装置]