A multi-focus image fusion method based on salient region extraction

Abstract

The present invention relates to a multi-focus image fusion algorithm based on salient region extraction, which comprises the following steps: 1. Using the GBVS algorithm to obtain saliency maps of two source images; 2. Binarizing the saliency maps and calculating their spatial frequencies respectively, Obtain the spatial frequency map; 3. Use the watershed algorithm to detect the weak edges of the spatial frequency map, obtain closed areas and classify them, and compare the spatial frequencies of different areas to find the pseudo-focused areas and filter them out, and use morphological methods for the remaining areas Obtain the focused area; 4. Directly fuse the focused area of ​​the source image, and decompose the defocused area through Shearlet transform to obtain its high-frequency and low-frequency sub-band coefficients, which are respectively fused based on PCNN and Laplace pyramid transform; 5. Reconstruct the fusion coefficient, Get a fused image. The gray level distribution of the fusion algorithm of the present invention is more scattered, the image texture is richer, and the details are prominent.

Description

technical field [0001] The invention relates to a method in the technical field of image processing, in particular to a multi-focus image fusion method based on salient region extraction. Background technique [0002] The limited focus range of the optical system makes it difficult to clearly present all objects in the scene, and the captured image generally consists of in-focus areas and out-of-focus areas. In order to obtain a clear and real scene, and obtain an all-focus fusion image containing feature information of different focus areas of the image, researchers continue to propose various multi-focus image fusion algorithms. [0003] According to the human visual attention mechanism, there are different salience between regions, and the focus of attention usually falls on the region with higher saliency. From a visual point of view, the focus area is a special salient area. In a multi-focus image, the focus area corresponds to the salient part of the image that attrac...

AI Technical Summary

Problems solved by technology

However, since the distribution of the salient area is related to the local contrast of the image, if the blurred area contains some areas with relatively high contrast, it will also be classified as a salient area in the saliency analysis, which will affect the judgment of the focused area.

Images