基于NSCT的红外与可见光图像融合方法

针对红外与可见光图像在融合时亮点目标易丢失且背景信息不够清晰的问题,提出基于非下采样轮廓波变换(NSCT)与局部区域融合规则相结合的红外与可见光图像融合方法。利用NSCT对源图像进行多尺度、多方向分解,对低频子带系数采用加权平均取均值的融合规则,对高频子带系数采用系数值选大与局部均方差相结合的融合规则,将前两步得到的系数进行逆NSCT得到融合图像。实验结果表明,在主观视觉效果和客观评价方面,该融合算法都是有效的且可得到比参考融合算法更为理想的融合图像。

Fusion method for infrared and visible light images based on NSCT

Aiming at the problem that the highlight targets are easy to lose and the background information is not enough clear during the fusion of the infrared and the visible light images,a fusion method based on the NSCT and local area fusion rules is proposed.The NSCT is used to take multi-scale,multi-direction decomposition for every source image.The mean of the weighted averaged coefficients is used to select the low frequency sub-band coefficients.As for the high frequency sub-band coefficients,a combination fusion method of choosing the maximum and local mean-square deviation fusion rules is developed.Finally,inverse NSCT is made to the coefficients obtained by the former two steps to get the fusion image.Experimental results show that the fusion algorithm is effective and can get better fusion image than the reference fusion methods both for subjective visual effect and objective assessment.