Research And Implementation Of Multiresolution Based Image Fusion Methods

When imaging a scene,due to the constraints of environment and the working mechanism of equipment,it is usually impossible to obtain a comprehensive description of the whole scene in a single picture.We can get a set of images by adjusting equipment parameters or using variant sensors,and then merge effective information by image fusion technology to enhance the scene description ability of hardware devices from software perspective.This paper improves image fusion methods,verifies the effectiveness of proposed method on the simulation software MATLAB,designs the image fusion system,and uses MFC+OpenCV to realize the fusion system.The research contents of this paper are as follows.We propose a method to fuse multiple exposure images.Image pyramid is used to process multiple images of different exposures in YCbCr space.In order to improve fusion color performance,the fusion strategies of brightness domain and chromaticity domain emphasize the proportion of chromatic aberration information.Experiments on 5 groups of multi-exposure images show that this method can balance details of light and dark areas in the scene and enhance the color performance of fusion image.We propose a method to fuse multiple focus images.Decomposing and reconstructing multiple images of different focus areas by SIDWT.The high frequency components are evaluated with regional energy consistency,and the local and regional features in images are comprehensively considered.The undecomposed images are distinct.Spectral Residual is used to calculate the saliance images of undecomposed images.Assign fusion weight to low frequency components with correlation between saliance images.Experiments on 12 groups of multi-focus images show that the method preserves information of focused area and improves the false contour of fusion image.We implement an image fusion system,with OpenCV and MFC,to achieve the proposed fusion methods.Image fusion system is mainly divided into image management module and image processing module.Image management module includes image read-in module,user read images from local document into the system;image display module displays image to be processed,the fusion result and its brightness histogram on the interface;Image storage module,user can save fusion result image to local document.Image processing module contains image fusion module,which takes multiple images read in by image read-in module as input data and calls fusion method to merge images.The fusion process is divided into three processes:decomposition,fusion and reconstruction.