摘要
以江苏省宜兴市为研究区,利用支持向量机(SVM)方法对资源三号测绘卫星影像进行了分类,其总分类精度为97.76%,Kappa精度为0.968 7。为了评价算法的适用性,同时应用最大似然法与最小距离法对同一影像进行分类测试,支持向量机分类法精度高于其他2种方法,可以满足土地覆盖分类调查需求。
Taking Yixing city as the research area,this paper classified ZY-3 satellite image using support vector machines.The classification result showed that total classification accuracy was 97.76%,total Kappa accuracy was 0.968 7.In order to evaluate the algorithm applicability,it calculated classification result by maximum likelihood method and minimum distance method.This study shows that SVM method has higher classification accuracy,and it can be used for land cover classification survey.
出处
《地理空间信息》
2013年第3期11-13,5,共3页
Geospatial Information
基金
国家科技支撑计划资助项目(2011BAB01B06)
关键词
资源三号测绘卫星
SVM
精度评价
ZY-3 Satellite,Support Vector Machines,accuracy evaluation