基于MODIS遥感数据的呼伦贝尔土地荒漠化信息提取  

Extraction of Desertification Information in Hulun Buir Based on MODIS Image Data

作　　者：孟翔冲[1] 姜琦刚[1] 齐霞[2] 王斌[1] 吴阳春[3] 李根军[4] 杨佳佳[1] 

机构地区：[1]吉林大学地球探测科学与技术学院,吉林长春130026 [2]沈阳地球物理勘察院,辽宁沈阳110121 [3]吉林大学综合信息矿产预测研究所,吉林长春130026 [4]吉林大学地球科学学院,吉林长春130026

出　　处：《安徽农业科学》2012年第5期2977-2979,共3页Journal of Anhui Agricultural Sciences

基　　金：中国地质调查局专项资助项目(1212010911084)

摘　　要：[目的]基于MODIS遥感数据,对呼伦贝尔地区土地荒漠化信息进行提取。[方法]基于空间分辨率为1 km的MODIS遥感数据,选取可以反映不同荒漠化特征的5个指标进行反演,并采用决策树分类方法对呼伦贝尔地区土地荒漠化信息进行提取。[结果]呼伦贝尔地区土地荒漠化面积为33 862 km2,占全区总面积的20.36%,且主要以沙质荒漠化为主;通过野外验证及高分辨率解译数据采点验证,此评价方法总体精度达89%以上。[结论]使用文中评价方法进行荒漠化监测,不仅能够对大尺度的荒漠化地区进行监测,而且具有较好的评价效果。[Objective] To extract desertification information of Hulun Buir region based on MODIS image data.[Method] Based on MODIS image data with the spatial resolution of 1 km,5 indicators which could reflect different desertification features were selected to conduct inversion.The desertification information of Hulun Buir region was extracted by decision tree classification.[Result] The desertification area of Hulun Buir region is 33 862 km2,accounting for 24% of the total area,and it is mainly dominated by sandiness desertification.Though field verification and mining point validation of high-resolution interpretation data,the overall accuracy of this evaluation is above 89%.[Conclusion] Evaluation method of this paper is not only effectively for large scale regional desertification monitoring but also has a better evaluation performance.

关 键 词：荒漠化 MODIS数据 遥感 决策树 反演 

分 类 号：S127[农业科学—农业基础科学]