Research On Multi-focus Image Fusion Based On Fuzzy Theory

As the types and application fields of imaging sensors continue to increase,multifocus image data obtained through them is increasing,and these data are both redundant and complementary.The use of redundant information can obtain more reliable results and improve the signal-to-noise ratio;the use of complementary information can make the fused image contain richer and more comprehensive details.Image fusion technology is the fusion processing of multiple source images of the same scene or target obtained by different imaging sensors to obtain clear,more comprehensive and reliable image information of the target according to actual needs.Image fusion technology is widely used in photography,computer vision,medical treatment,remote sensing technology,aviation and other fields.It has high application value,so it is of great significance to the research of image fusion.The reason why image fusion technology is used is to extract images to obtain clear areas,and then combine these clear areas to obtain clear images.Therefore,it is clear that the fusion quality of the multi-focus image will affect the clarity of the entire image.Fuzzy theory can describe the uncertainty in the image well,and it is one of the important tools to study the ambiguity and uncertainty.This paper applies it to multi-focus image fusion,and proposes two multi-focus image fusion algorithms based on fuzzy theory:(1)A multi-focus image fusion algorithm based on fuzzy reasoning is proposed.Firstly,perform fuzzy inference on the source image and obtain the focus metric map of the source image;secondly,compare the obtained focus metric maps to obtain the fusion decision map;finally,fuse the two multi-focus source images according to the obtained decision map.(2)A multi-focus image fusion algorithm based on non-subsampled shearlet transform(NSST)and type-2 fuzzy is proposed.Firstly,non-subsampled shearlet decomposition is performed on the two source images separately,and the high-frequency subbands after decomposing are enlarged to better preserve image details;low-frequency subbands are fused using type-2 fuzzy to make the transition of the image edges more natural.Finally,the inverse transform is performed on the high and low frequency subbands to generate the fusion result map.The two fusion algorithms are respectively simulated and compared with three classical fusion algorithms.The comparison of subjective and objective fusion evaluation criteria validates the effectiveness of the multi-focus image fusion algorithm based on fuzzy theory.