A Multi-Focus Image Fusion Algorithm Based on Weighted Multi-Scale Morphological Gradients and Uniform Segmentation

Abstract

Multi-focus image fusion is a technique used to create a single image that is entirely in focus by merging the focused regions from multiple images. This paper presents an approach to multi-focus image fusion that combines uniform segmentation with a weighted multi-scale morphological gradient focus measurement. During the creation of the focus degree measurement map of the source image, multi-scale morphological gradients are first obtained through multi-scale morphological operations. These gradients are then adjusted using scale weighting and distance weighting. For image segmentation, to minimize the occurrence of image blocks containing both focused and defocused pixels, the image is uniformly divided into 2x2 pixel blocks. The focus degree value of each block is determined by the values within a 6x6 pixel range that includes the block and its surrounding pixels. Experimental results demonstrate that this method enhances the accuracy of focus degree measurement, particularly in smooth regions of the image.