结合纹理信息和支持向量机的金钟藤信息提取  

Merremia boisiana information extraction based on support vector machine method combined with texture features

作　　者：姜圆圆[1] 王正海[1,2] 

机构地区：[1]中山大学地球科学与地质工程学院,广州510275 [2]广东省地质过程与矿产资源探查重点实验室,广州510275

出　　处：《测绘科学》2016年第5期100-104,共5页Science of Surveying and Mapping

基　　金：广东省教育科技创新项目(2013KJCX0008)

摘　　要：为了寻找快速准确、方便调查病态性入侵植物金钟藤分布的方法和技术,该文以海南省定安县为研究区,结合光谱信息与纹理信息,运用支持向量机的分类方法,提取该地区金钟藤信息,并将分类的结果与最大似然法、单源数据(B3波段图像)的支持向量机分类结果进行定性与定量的比较分析。结果表明:基于比值植被指数、B3波段为纹理信息源的支持向量机方法有利于金钟藤信息的提取;通过结合纹理信息和支持向量机的方法,实现了分类精度的提高。In order to efficiently monitoring and analyzing the distribution of the compulsive invasive plant merremia boisiana,the paper taking Ding＇an Country of Hainan Province as the example,used the classification method of Support Vector Machine to extract the distribution of Merremia boisiana combining with the spectral and textural information.Moreover,the classified result was comparatively analyzed with Maximum Likelihood method and the Support Vector Machine method with single-source data in a qualitative and quantitative way.Results showed that the method combining the texture information and Support Vector Machine would be better for extracting the information of Merremia boisiana with higher classification accuracy.

关 键 词：金钟藤 支持向量机 纹理信息 植被指数 

分 类 号：P237.9[天文地球—摄影测量与遥感]