研究山区地表水体信息OLI遥感数据去阴影自动提取方法,设计基于数字高程模型与指数提取的决策树分类方法,提高水体自动识别的精度.该方法选取改进的归一化水体指数、归一化植被指数、比值植被指数、主成分分析前3个分量以及波段之间的组合运算,并结合DEM构建决策树分类规则.综合采用单波段阈值、谱间关系、植被指数和水体指数阈值完成山体水体的去阴影识别研究,与计算机自动识别分类方法比较,其精度明显提高.结果表明,决策树分类方法在精度上明显高于常用的计算机自动分类方法,可以很好地被利用于OLI遥感数据水体信息的海量、大范围提取.%The automatic extraction method of OLI remote sensing data for surface water body in mountainous area is studied.The decision tree classification method based on DEM and index extraction is designed, which improves the accuracy of automatic identification of water body.The improved normalized water index (MNDWI), ratio Vegetation Index (RVI) and normalized difference vegetation index (NDVI) are selected, and the decision tree classification rules are constructed with the digital elevation model (DEM).At the same time, the method of threshold value method, spectrum-photometric method, vegetation index and water index threshold method are used.The research result shows that the decision tree classification method is obviously higher than other methods, which can be used in the extraction of surface water of OLI image data.
展开▼
