Multi-focus Image Fusion Algorithm Based On Guided Filtering

Multi-focus image fusion(from here on referred to as MFIF)effectively combines the information of the same scene in different focal regions,so that objects at different imaging distances can be clearly presented in one image,which solves the defects of limited focus range of optical imaging systems.MFIF allows people to provide more accurate,comprehensive and reliable information.This lays a good foundation for target recognition,feature extraction and tracking detection.MFIF is widely used in the fields of digital photography,computer visual recognition,image segmentation,and microscopic imaging.It also has practical significance and application value.In addition,the guided filter is a novel edge-preserving filter that has a good edge smoothing function as well as an effective edge-gradient-preserving characteristic,and low complexity.For the problems of related to multi-focus image fusion algorithms,this paper proposes two multi-focus image fusion algorithms based on guided filtering:(1)Research on a multi-focus image fusion algorithm based on nonsubsampled contourlet transform and guided filteringIn order to solve the problem of virtual shadows appearing at the edges of the target objects in multi-focus image fusion,a multi-focus image fusion algorithm based on non-subsampled contourlet transform and guided filtering is proposed.Firstly,multi-scale decomposition of multi-focus source images are performed by a non-subsampled contourlet transform.Then,in order to enrich the edge integrity of the multi-focus fused image,a guided filtering-weighted fusion rule,based on edge integrity,is proposed in the low frequency sub-band coefficient.A guided filter-weighted fusion rule,based on significant information,is proposed in the band-pass direction sub-band coefficients.Finally,a non-subsampled contourlet reconstruction is performed to obtain a multi-focus fusion image.The final simulation results show that the proposed algorithm can effectively extract edge information,reduce artifacts,and improve the quality of fused images.(2)Research on a multi-focus image fusion algorithm based on nonsubsampled shearlet transform and guided filteringIn order to further improve the sharpness and contrast of the multi-focus fused image and accord with the human visual characteristics,a multi-focus image fusion algorithm based on non-subsampled shearlet transform and guided filtering is proposed.Firstly,multi-scale decomposition of multi-focus source images are performed by using non-subsampled shearlet transform.Then,in order to improve the sharpness of the multi-focus fused image,a guided filtering-weighted fusion rule,based on local region improved spatial frequency,is designed in the low-frequency sub-band coefficient.In order to accord with the human visual characteristics,a guided filter-weighted fusion rule,based on human vision features,is used in the band-pass direction sub-band coefficients.Finally,non-subsampled shearlet reconstruction is used to obtain multi-focus fusion image.The final simulation results show that the proposed algorithm further improves the contrast and sharpness of the fused image,and more accord with human visual characteristics.It is an effective multi-focus image fusion algorithm.In order to analyze and evaluate the performance of the above two image fusion algorithms efficiently and intuitively,the other nine image fusion algorithms are compared with the two image fusion algorithms.A multi-focus image fusion software system was integrated and designed in order to complete this research.The implementation of the software system provides technical support for the application of multi-focus image fusion.