基于SVM的湖泊咸度等级遥感信息提取方法——以内蒙古巴丹吉林沙漠为例  

Extraction of remote sensing information for lake salinity level based on SVM:A case from Badain Jaran desert in Inner Mongolia

作　　者：刁淑娟[1] 刘春玲[2] 张涛[2] 贺鹏[2] 郭兆成[2] 涂杰楠[2] 

机构地区：[1]中国地质图书馆,北京100083 [2]中国国土资源航空物探遥感中心,北京100083

出　　处：《国土资源遥感》2016年第4期114-118,共5页Remote Sensing for Land & Resources

基　　金：中国地质调查局地质调查项目"巴丹吉林沙漠1:25万水文地质调查"(编号:12120114018501)及"区域水文地质遥感调查技术要求(1:50 000;1:250 000)"(编号:121200114029401)共同资助

摘　　要：以Landsat8 OLI图像为信息源,在完成几何纠正、配准及镶嵌等图像预处理的基础上,对内蒙古巴丹吉林沙漠湖泊进行光谱信息提取;结合实地采集的湖泊咸度等级样本,分析湖泊咸度光谱特征及其遥感信息特征,提出了基于支持向量机(support vector machine,SVM)的咸度等级分类方法和咸度区划分方案,实现了对巴丹吉林沙漠湖泊咸度等级遥感信息的提取。研究结果表明,该方法能有效地解决湖泊咸度样本数量较少、样本间光谱信息差异较小、采用常规分类方法难以取得较好效果情况下的湖泊咸度信息提取问题,有一定的推广应用价值。According to the problems of remote sensing information extraction in lake salinity level of Badain Jaran desert, the authors put forward a method based on support vector machine （SVM）. In this paper, the authors adopted Landsat8 OLI remote sensing image as the data source, completed the image preprocessing such as geometric correction, image registration and mosaicking. With the help of the RS and GIS technology, the authors successfully extracted the information of lake salinity levels of the Badain Jaran desert. The results show that the proposed method can effectively solve the problems of less samples and the information extraction of lake salinity levels when the spectral information is confused, and hence has the reference value and can be promoted to other similar situations.

关 键 词：湖泊咸度等级 信息提取 支持向量机(SVM) 光谱分析 巴丹吉林 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]