青藏高原深切割区GF-1土地利用分类精度研究  

Research on Land Use Classification Precision of GF-1 Images in the Deep Incised Area of Qinghai-Tibet Plateau

作　　者：孙小飞[1] 刘智[2] 范敏[2] 韩磊[2] 

机构地区：[1]成都理工大学地学空间信息技术国土资源部重点实验室,四川成都610059 [2]四川省地质调查院,四川成都610081

出　　处：《地理空间信息》2017年第10期13-15,共3页Geospatial Information

基　　金：国家自然科学基金资助项目(41302282;41401659);全国边海防地区基础地质遥感调查资助项目(12120115063501);四川省科技厅应用基础资助项目(2015JY0145)

摘　　要：以青藏高原深切割地区为例,研究高分一号影像在深切割地区的分类方法。分别采用最大似然法、神经网络法和支持向量机分类法对影像进行分类,并对分类精度进行评价和分析。结果表明:1)高分一号区分地物的精度较高,可满足土地利用的遥感监测;2)支持向量机分类方法其总体精度为91.67%,Kappa系数为89.54%,均高于最大似然法和神经网络分类法。因此,支持向量机分类方法更适合于高分一号影像在深切割地区的分类。Taking the deep incised area of Qinghai-Tibet plateau for example, this paper studied classification technique and methods in the deep incised area of GF-1 images. The paper used the maximum likelihood method, the neural network method and the support vector machine(SVM) classification method to classify the images, evaluated and analyzed the precision of three classification methods. The results show that(1)?the precision of classification by using GF-1 images is high, which meets the requirement of remote sensing monitoring of land use.(2)?The whole precision of SVM method is 91.67% and it's Kappa coefficient is 89.54%, which is more precise than the maximum likelihood method and the neural network method. In conclusion, SVM method is optimal one to classify land use in deep incised area by using GF-1 images.

关 键 词：深切割区 高分一号 分类 精度 

分 类 号：P237.9[天文地球—摄影测量与遥感]