基于多源遥感的土地利用动态监测图像分类方法研究——以陕北黄土丘陵沟壑区为例  

Image Classification Method in Landuse Dynamic Detection Based on Multi-source Remote Sensing Data——A Case Study in the Loess Plateau of Northern Shaanxi Province

作　　者：刘咏梅[1] 李锐[2] 杨勤科[2] 

机构地区：[1]西北大学城市与资源学系,陕西西安710069 [2]中国科学院水利部水土保持研究所,陕西杨凌712100

出　　处：《水土保持通报》2006年第6期63-66,共4页Bulletin of Soil and Water Conservation

基　　金：黄委会项目第四专题"区域水土流失模型试验研究"(2004SZ01-04);"十一五"国家科技支撑计划项目"水土流失动态监测与评价关键技术"(2006BAD09B05)

摘　　要：在陕北黄土丘陵沟壑区的土地利用动态监测中,采用一种遥感影像和单纯的监督分类方法,难以获得高精度的土地利用数据。为解决此问题,以陕北无定河流域为研究区,以主成分变换的方法对多源遥感影像(TM多光谱数据和SPOT全色波段数据)进行融合处理;同时在分类中采用监督分类与非监督分类相结合的混合分类法,改进训练样本选取方法。先以非监督分类获得初始训练样本,在对样本进行删除、增补、合并等调整的基础上,再进行监督分类,这2种方法的结合使用,使土地利用信息自动提取的精度明显提高。与仅以TM影像为信息源,采用单纯监督分类法的分类结果对比可知,土地利用各类别的提取精度都有不同程度的提高,分类总精度从82.0%提高到89.2%;水体、水田和城镇用地等面积较小类别的精度提高了10%以上;坡耕地与林草地的混分现象明显减少,精度均提高了5%以上,取得了良好的分类效果。研究结果为陕北黄土丘陵沟壑区土地利用变化的动态监测,提供了重要的技术支持和借鉴。The landuse classification accuracy based on single remotely sensed data and simple supervised classification is unsatisfactory to the landuse investigation in loess hill and gully area. By taking the Wuding River watershed in Northern Shaanxi Province as a test area, the TM muhi-spectral data and SPOT pan data were merged by the method of Principal Components Analysis. Based on the merged image, the landuse categories were then extracted by applying an integration of supervised classification and unsupervised classification. The combination ot two methods remarkably improved sampling method. Compared to the classification based on the single TM multi-spectral data and supervised classification, the total accuracy increased from 82.0 % to 89.2 %, especially the accuracy of city and town area, paddy field, water area increased over 10%, the mixture of sloping land and forest （grass-land） decreased remarkably, and the accuracy of the two categories increased over 5 %. The result is of critical significanee in landuse dynamic monitoring in the area.

关 键 词：影像融合 土地利用 分类 陕北黄土丘陵沟壑区 

分 类 号：P237[天文地球—摄影测量与遥感] S157[天文地球—测绘科学与技术]