基于多级分块迭代法的不同聚焦图像融合

Multi-Focus Image Fusion Based on Multi-Level and Iterative Method

提出了一种基于多级分块迭代法的不同聚焦图像融合方法.首先,通过采用高斯函数对图像进行多次卷积以模拟图像的再次成像系统,对原始图像和再次成像采用不同的块大小进行多级划分,在不同的级别上比较原始图像和再次成像的灰度差平均值判断子块清晰与否.接着,根据一致性检验原则对图像块进行隶属情况判断.然后,采用迭代法将图像划分成三个区域:清晰区域、模糊区域、交界区域.最后,选择其中的清晰区域形成融合图像.文中给出了融合实验结果,结果表明该方法与传统融合方法相比性能优越,能够大幅度提高融合图像的质量.

A new image fusion method for multi-focus image is presented based on multi-level sub-block and iterative method.In order to simulate the system of re-imaging,multi-focus images are convolved with Gaussian functions.Source images and re-imaging images are partitioned into multi-level sub-blocks by different sizes.The gray deviation of the average between source images and re-images is used to judge whether the block is focused.The subjection of image blocks are judged by the principle of consistency.Images are divided into three regions：clear region,fuzzy region and border region,based on iteration method.Finally,the clear regions are merged into the fused image.Experimental results show that the proposed method provides superior performance over conventional fusion methods in improving the quality of fused images.