基于地物特征的遥感图像融合效果

Fusion effect of remote sensing images based on object features

以ETM多谱段和全色影像为数据源,在最佳波段组合分析的基础上,采用主成分变换、乘积变换、小波变换、小波+IHS混合变换和小波+主成分混合变换等方法对图像进行融合.以均值、标准差、信息熵为评价指标,比较分析了不同地物特征的图像融合效果.结果表明:采用小波+IHS混合方法有利于耕地信息的提取;采用小波+主成分融合方法有利于居民地、林地、牧草地、水域、未利用地和园地等信息的提取.

Using the best wave-band combinations with ETM and Pan images based on object features,the fusion effect of image was compared and analyzed by the methods of principal components transformation fusion,multiplicative transformation fusion,wavelet transformation fusion,wavelet＋IHS and wavelet＋principal components transformation fusion.The results were as follows： the fusion result of wavelet＋IHS was favorable for the extraction of farmland information;the fusion result of wavelet＋main components was favorable for the extraction of habitation,forest lands,pastures,water areas,unutilized lands and garden plots.