Research on Multi-Focus Image Fusion Algorithm Based on Bilateral Weighted Guided Filtering in Pyramid Domain

为了克服传统多聚焦图像融合方法中的边缘失真和图像模糊问题，提出了一种金字塔域双边加权引导滤波的融合方法。首先，利用双边加权引导滤波对均值滤波和差分算子得到粗糙焦点图进行细化并作为金字塔融合过程中高频信息选择策略，在金字塔下采样过程中，采用双边加权引导滤波替代传统的高斯滤波，更有效地保留图像的边缘信息。同时，为了提升图像细节部分，以原始待融合图像的均值作为参考，进行差异运算，构建了两个图像金字塔，并对每个金字塔单独处理，之后利用高频信息决策图进行叠加，通过重建金字塔得到融合图像。实验结果表明，无论是在主观视觉感知还是在客观评价指标方面，本文所提出的方法在视觉保真度等多个指标上优于经典方法，具有较好的图像融合效果。In order to overcome the problems of edge distortion and image blur in the traditional multi-focus image fusion method, a fusion method based on Bilateral Weighted guided filtering in the pyramid domain was proposed. Firstly, the bilateral weighted guided filter is used to refine the rough focus map obtained by the mean filter and the difference operator, and as a high-frequency information selection strategy in the pyramid fusion process, the double-sided weighted guided filter is used to replace the traditional Gaussian filter in the pyramid downsampling process, so as to retain the edge information of the image more effectively. At the same time, in order to improve the detail of the image, the average value of two original images with a different focus was used as a reference, and two image pyramids were constructed by different operations, and each pyramid was processed separately, and then the high-frequency information decision map was used to superimpose and obtain the fusion image by reconstructing the pyramid. Experimental results show that the proposed method is superior to the classical method in terms of visual fidelity and other indicators, and has a better image fusion effect.