Image Fusion Algorithm Based On Region Energy And Block Matching

With the development of image sensor and high speed processor,the image information processing technology has been developed.Image fusion is an important branch of the image processing field.Image fusion is the image processing technology that makes fuse different images into one complete,collecting the multi image information of the same scene through the multiple optical sensors.And the image fusion technology can get image information accurately and comprehensively,improving the utilization rate of the image information which is collected by multi sensors.And this technology can overcome the limitations of the single image.Therefore,the image fusion technology can be widely used in the fields of meteorology,military,medicine,artificial intelligence and so on.The image fusion algorithm can be divided into two categories:spatial domain image fusion algorithm and transform domain image fusion algorithm.The spatial domain algorithm is simple and intuitive,and the processing is efficiency.The results of the transform and fusion algorithm are much better.Based on the BM3D(Block Matching 3 Dimension)denoising algorithm,the algorithm in this thesis combines the spatial domain processing technology and transform domain information processing technology and fully uses the known information.At the same time,a complete image fusion system is finally formed based on the regional correlation between the regional energy and the mean gradient,and the selection of fusion rule which is based on matching degree.The main contents of this thesis are as follows:(1)This thesis contains the traditional framework of classical algorithm in pixel level image fusion and the latest achievements in the image fusion field.And the method in evaluating the image and evaluating indicator have been studied and mastered,and select the method of the optimal evaluating indicator according to the specific experimental requirements;(2)This thesis contains the basic framework of the BM3D algorithm,as well as the specific steps and algorithms in the framework;(3)The weighted idea and average gradient are introduced into the framework of image fusion rules.After studying some related theories,this thesis presents a new method of image fusion that the image representation coefficients are divided by the matching degree.And then,this method deal thedifferent image representation coefficientsaccording to the different fusion rulesas matching;This paper puts forward a new framework of image fusion algorithm that is based on the improved BM3D algorithm and the improved region energy algorithm based on matching degree,Summarize the theoretical research results and shortcomings of this thesis,and prospects the future research.