SPOT-5数据在退耕还林监测中的应用研究——以山西省古交市嘉乐泉乡为例  

Study on Monitoring Conversion of Cropland to Forest by Using Spot-5 Data——A Case study in Jialequan Town,Gujiao County,Shanxi Province

作　　者：史良树[1] 杜纪山[2] 谭炳香[3] 

机构地区：[1]中国土地勘测规划院,北京100035 [2]国家林业局退耕还林办公室,北京100714 [3]中国林业科学研究院资源信息研究所,北京100091

出　　处：《林业资源管理》2006年第6期84-87,共4页Forest Resources Management

摘　　要：以山西省古交市嘉乐泉乡为试验区,采用SPOT-5的10m、5m和2.5m 3种影像数据对退耕还林地面积进行分类监测。所设计的2种方案分别是:1)将地物类型分为7类,退耕还林地作为一种单独地类,对3种影像数据进行计算机自动分类和2.5m影像的人工解译分类;2)借助退耕还林作业设计图,将退耕还林地块影像分割出来,对退耕还林地和未退耕还林地进行有监分类。精度验证表明,第一种方案中2.5m融合图像的人工解译分类,退耕还林地的分类精度在50%以下;第二种方案中3种影像数据的总体分类精度均大于90%。建议在退耕还林地的作业设计图电子化的基础上,应用SPOT-5数据监测退耕还林地的任务完成和植被覆盖情况。Three kinds of SPOT- 5 images with different spatial resolution of 10m, 5m and 2.5m were used to monitor the area of CCF（conversion of cropland to forest） RFTF in Gujiao County, Shanxi Province. There are two schemes adopted in the monitoring classification of images： 1 ） seven landuse types were adopted and the land of CCF was regarded as one of them. The automatic classification method was used in the three images respectively and visual interpretation was used for the 2.5m image; 2） the image of the area of CCF was extracted from the image of the study area based on electronic maps of afforestation design of CCF. Then, two classification types- CCF land and non CCF landand supervised classification are adopted. The results have showed that the first scheme has a classification accuracy of less 50 % for visual interpretation of 2.5m image through analyzing the classification, while the second scheme has a classification accuracy of over 90% for all images. It is suggested that SPOT - 5 image can monitor the finishing area and plant coverage of CCF by using electronic maps of afforestation design of CCF.

关 键 词：SPOT-5影像 退耕还林 图像融合 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]