基于NSCT和ICA的红外和可见光图像融合方法

Fusion technique for infrared and visible light images based on independent component analysis and non-subsampled contourlet transform

针对基于非下采样轮廓波变换(non-subsampled contourlet transform,NSCT)的红外和可见光图像融合存在的目标信息不明确、融合图像对比度较低的问题,提出了NSCT和独立分量分析(independent component analysis,ICA)的红外和可见光图像融合方法。首先采用NSCT对红外和可见光图像进行多尺度、多方向分解,然后对分解后的红外和可见光图像的低通子带系数采用基于ICA的图像融合方法,得到低通融合图像。再使用以邻域系数差和信息熵为标准的带通图像融合规则对带通子带系数进行融合,得到带通融合图像。最后对低通融合图像和带通融合图像进行NSCT的逆变换,从而得到最终的融合图像。仿真实验验证了本文方法的有效性。

The infrared and visible light image fusion based on the non-subsampled contourlet transform （NSCT） can lead the target information to be uncertain and causes the problem of low contrast. As to the issue, a new algorithm by combining the NSCT with the independent component analysis （ICA） is presented. Firstly, the NSCT is applied to the infrared and visible light image for multi-scale and multi direction decomposition. Then, the low-pass sub-band coefficients in the decomposed images are fused by ICA to attain the low-pass fu sion image. Meanwhile, the band-pass sub band coefficients are fused by the band-pass image fusion rules with neighborhood coefficient difference and information entropy as criteria. Finally, the inverse transform of NSCT is used to fuse the low-pass fusion image and the band-pass fusion image to gain the final fusion image. The sim- ulation validates the efficiency of the proposed algorithm.