基于Iαβ色彩空间和Contourlet变换相结合的融合方法

Image fusion technique based on combination of Iαβ color space and Contourlet transform

针对北京1号小卫星的多光谱与全色波段的分辨率比率较大,传统的融合方法会产生边界模糊和光谱扭曲现象,提出了一种新的融合算法。首先对多光谱与全色影像分别进行Iαβ和Contourlet变换;然后在频率域中采用不同的融合策略进行处理;最后进行Contourlet和Iαβ逆变换,得到融合图像。实验表明,本方法既提高了融合图像的空间细节信息又很好地保持了图像的光谱特征,优于传统的融合方法。

With the relatively higher resolution ratio of panchromatic and multispectral images of Beijing-1 Satellite,blurred boundaries and spectral distortion always occur when they are fused employing traditional fusion approaches.To resolve this problem,proposed a new fusion technique.Firstly,transformed multispectral and panchromatic images by Iαβ and contourlet,and then processed with different fusion schemes in frequency domain.Finally,through inverse Iαβ and contourlet transforms,worked out the fused image.The test results indicate that this fusion approach enhances the spatial detail information of the fused image and preserves the spectral characteristics of original image,which is better than traditional fused methods.