结合方向信息测度和NSCT分解的图像融合算法

Image Fusion Algorithm Combined with Orientation Information Measure and NSCT

针对传统图像融合算法容易造成边缘细节信息丢失且抗噪声能力不强的问题,提出一种结合方向信息测度和非下采样轮廓波变换分解的图像融合算法,先利用方向信息测度对图像进行复原操作,再对低频部分采用方向信息测度加权融合规则进行融合,对高频部分采用改进区域能量加权融合规则进行融合.该算法较好地结合了非下采样轮廓波变换良好的信息保持能力和方向信息测度良好的抗噪性能,使得融合图像的效果良好.实验证明,该算法能较好的保持边缘信息且抗噪性强,能广泛地应用于多聚焦图像的融合.

It easily happened that the edge information lost and the effect of anti-noise was weak if the traditional image fusion algorithms were applied. In order to preserve the edge information in the source image and improve the noise immunity, a new image fusion algorithm based on orientation information measure and NSCT was presented. The algorithm firstly recovered source image by using Orientation information measure, and then performed it as low frequency and high frequency parts by using NSCT decomposition, finally, fused the low frequency part by using Orientation information measure weighted rule, and the high frequency part by using this paper＇s improved region energy weighted rule. The algorithm maked full use of the good information retention ability with NSCT and the good anti-noise performance with orientation information measure, such that the fused image had good results. Experiments show that, this algorithm can preserve the edge information with strong noise resistance, can be widely applied in multi-focus image fusion.