耕地变化的影像比例化分析  

Analysis of Image Pixel Proportion on Cultivated Land Change

作　　者：李向军[1] 牛铮[1] 王汶[1] 吴运超[1] 

机构地区：[1]中国科学院遥感应用研究所遥感科学国家重点实验室,北京100101

出　　处：《遥感信息》2005年第4期38-40,i0001,共4页Remote Sensing Information

基　　金：中国科学院知识创新工程重大项目(KZCX1-SW-01-02);国家高技术研究发展计划(863计划2003AA131170)共同资助。

摘　　要：应用遥感手段检测耕地变化常采用分类后处理方法,这种方法在分类标识过程中没有充分考虑混合像元因素影响。针对这一现象,本文引入纹理分析做法,对分类后影像实施划分一致性区域和确定零星像素归属,最后各像素以比例形式反映耕地的变换。The post-classification is one of the change detection methods in remote sensing and it is often used to detect cultivated land change. But the traditional approach does not fully consider the effects caused by mixed pixcels in the classified results. To decrease this effect, the article makes an improvement to this approach. We take advantage of the approach of texture analysis, and make further subdivision of post-classification images. Finally we could reflect cultivated land change in the form of proportion of pixels.

关 键 词：耕地变化 覆盖比例 遥感 

分 类 号：P237.9[天文地球—摄影测量与遥感]