基于SVR和贝叶斯方法的全色与多光谱图像融合

将全色图像和多光谱图像进行融合,可以获得高空间分辨率和高光谱分辨率的融合图像.利用支持向量回归（SVR）模型构建的支持向量值轮廓波变换,对源图像进行多尺度、多方向、多分辨率分解；采用贝叶斯方法获得在不同分解水平上的全色图像和多光谱图像融合算法；利用支持向量回归的强大学习能力,通过全色图像和多光谱图像之间的相关关系,获得超分辨率的多光谱图像,解决贝叶斯方法中的待融合图像分辨率一致性问题.实验结果表明,采用该方法获得的融合图像既具有较高的空间细节表现能力,又保留了多光谱图像的光谱特征,融合效果优于传统的图像融合方法．

Fusion of panchromatic image and multi-spectral image based on SVR and Bayesian method

The fusion images with high spatial resolution and high spectral resolution can be obtained by fusing panchromatic images and multi-spectral images. Support vector value contourlet transform constructed by using support vector regression model was used to decompose source images at multi-scale, multi-direction and multi-resolution. The algorithm of fusing panchromatic image and multi-spectral image was derived at different levels by using Bayesian method. By utilizing the strong learning ability of support vector regression and the relationship of multi-spectral image with panchromatic image, the super-resolved multi-spectral image was reconstructed to resolve the problem of coincident resolution of images to be fused. Experimental results show that the fused image obtained by the method not only has high spatial resolution, but also preserves the spectral characteristics of the multi-spectral images. The fusion performance of the method is better than traditional image fusion methods．