基于小波的遥感影像特征信息融合模式研究

在数字图像小波多分辨率分析的理论基础上,采用小波变换方法对高分辨遥感图像的目标地物边缘进行信息增强,然后与多光谱遥感图像进行特征信息融合.在融合过程中,首先对多光谱图像中的R、G、B 3个波段的图像进行小波分解,得到相应的低频图像,并对特征增强后的高分辨率图像进行小波分解,再将分解后的高频图像分别与低频图像进行融合,然后经RGB合成为彩色图像.该方法既改善了图像的清晰度和分辨率,同时也保留了原图像的光谱信息.最后通过融合实验验证了上述结论.

Wavelet Based Fusion Model of Feature Information in Remote Sensing Images

Fusion of feature information in remote sensing (RS) images is generally based upon spectral and texture features in object images. This paper presents a feature fusion approach to RS images after enhancement of edges in the remote sensing image with multi-resolution wavelet analysis. The image resolution is improved. Instead of the conventional HIS-wavelet fusion, fusion processes of the low frequent components of the RGB channels and the high frequency components, both obtained with wavelet decomposition, are performed separately, and a new fusion image is composed of the new RGB channels. In this way, not only resolution of the multi-spectral image is improved, but also color features, which are important for multi-spectral fusion, are retained. Effectiveness of the proposed method is verified with experimental results.