Research On Filtering Based Multi-focus Image Fusion Algorithm

Multi-focus image fusion is an important research area in the field of image fusion,which can overcome the limited depth of field of the optical lens,extract clear areas in different images with different focus ranges,and clearly display the extracted areas in an image,which is beneficial to feature extraction,edge detection and target recognition of an image.Multi-focus image fusion involves many disciplines such as sensors,signal processing,image processing,etc.It has been widely used in medical diagnosis,remote sensing,surveillance and military fields.The multi-focus image pixel level fusion is the basis of multi-focus image fusion.How to effectively extract the focus area in the source image and obtain more edge detail information is the key to multi-focus image pixel level fusion.The purpose of this paper is to improve the accuracy of the focus judgment of the source image,effectively suppress the block effect and to better extract the clear region and the details of the edge target.The main research works of this paper are as follows:1.It is difficult to accurately determine the pixels in the focus area and the out-of-focus area.Besides,the edge contour and detail information may lose during the fusion,a multi-focus image fusion method based on guided filtering(GF)is proposed.By using GF to layer the source image,the basic components and detailed components of the source image are obtained to determine the focus area of each component to obtain the fused image.This method can effectively identify the clear regions in the source image,and produce clearer details and boundaries of the fused image.2.For non-negative matrix factorization(NMF)algorithms,the fused image are of low blurring and low contrast.A multi-focus image fusion algorithm based on dynamic guided filtering(RGF)is proposed.The multi-scale decomposition of the fused image is used to obtain the base layer and the detail layer.Different fusion algorithms are used for each layer of the source image to make the texture and structure information of the fused image clearer and the contrast is significantly improved.3.The fusion algorithms are integrated into the image fusion system.A multi-focus image fusion prototype system is designed and implemented.