土地利用的遥感识别方法研究

土地利用状况是区域规划、农作物合理布局、生态环境保护、流域水文模型建立、防洪救灾等的重要依据。本文探讨了利用不同时相陆地卫星TM信息及其植被指数(NDVI)识别土地利用类型的四种方法：①TM最佳波段组合的彩色合成解译法；②TM最佳波段组合的混合聚类法；③不同时相NDW.图的彩色合成解译法；④不同时相NDVI图的混合聚类法。通过在安徽省滁县试验区的应用表明,四种方法联合使用可有效地识别土地利用类型。

LAND USE SURVEY USING REMOTE SENSING DATA:A METHODOLOGICAL STUDY

The state of land use ban provide important basis for regional plan, crops layout,ecological environment protection,basin hydrological model establishment and flood control, etc. Four methods of land use survey using multiseasonal multispectral Landsat TMdata and the normalized difference vegetation index (NDVI) calculated from Landsat TMdata are discussed,in this paper. They are ①interpretation of TMfalse colour composite image composed by three optimal band TMimages; ②mixed cluster classification (MCC) using optimal band TMimages; ③interpretation of NDVIfalse colour image composed by optimal seasonal NPVIimages ; ④MCCusing optimal seasonal NDVIimages. Applied these methods to land use survey in an experimental area of Chu County, Anhui province, it shows that land use types can be surveyed efficiently by using these four methods in the same time.