针对SAR与光学图像的融合问题，提出一种基于SAR图像中纹理特征的Contourlet变换融合方法。利用灰度共生矩阵法提取SAR图像的纹理特征，分析各个纹理特征间的相关性，得到重要纹理特征图。用 HSV变换提取光学图像的强度分量。将重要纹理特征和强度分量利用改进的Contourlet多尺度变换融合，得到新的强度分量。通过HSV逆变换得到SAR与光学的融合图像。利用Landsat8和Cosmo-SkyMed图像进行融合实验，并与小波、HSV、Brovey、Contourlet变换融合方法对比分析，实验表明该方法能够较好的保持光学图像的光谱特征和SAR图像的纹理、强散射特征，增加图像细节信息，提高图像可解译性。%For SAR and optical image fusion problem ,a fusion method based on Contourlet transform for texture characteristics is proposed .Firstly ,texture features are extracted from gray level co-occurrence matrix in SAR image ,then the main texture characteristics are obtained from important texture features by analyzing the correlation between the texture characteristics .Secondly ,intensity component of optical image is extracted through HSV transform algorithm .The new intensity component is given by the use of the fusion method of improved Contourlet multi-scale transform .Lastly ,the fusion of SAR and optical images are finished through HSV inverse transformation .The Landsat8 and Cosmo-SkyMed images are selected as the fusion experiments .And compared with wavelet ,HSV ,Brovey ,Contourlet fusion method , the experiment results show that the texture and strong scattering characteristics of SAR image and optical image spectrum characteristics are kept better ,in w hich the image detail information is increased ,and beneficial to image interpretation .
展开▼
