【摘要】 基于SPOT-5多光谱和全色影像进行图像融合,构造NDVI植被指数,提取多尺度纹理特征并参与分类;基于SVM方法进行森林资源植被分类.结果表明:高通滤波(HPF)为最佳的图像融合方法,与单尺度纹理与光谱信息融合的分类相比,多尺度纹理与光谱信息融合的分类方法总精度提高了3%,验证了新方法的可行性.更多还原

【Abstract】 The images were fused,the NDVI vegetation index was structured and the multi-scale texture characteristics were extracted for forest resources vegetation classification with SVM method based on SPOT-5multispectral and panchromatic images. The results showed: HPF was the best images fusion method,the classification accuracy with multi-scale texture and fused spectral information increased 3% comparing with the method of single scale texture and fused spectral information,the feasibility of new method is verified.更多还原