Study On The Images Fusion Algorithm Based On Wavelet Transform

Image is developed from the information fusion technology. It can beadvantageous to extract the maximum information of each channel, and finallyintegrated into high-quality images to improve the utilization of the image information.The fused image not only has the most information of the source image, but alsoconsistent with human visual characteristics and machine perception, which makes theimage fusion technology has been widely applied to become a hot topic today. Withthe increasing maturity of the wavelet transform, wavelet multi-scale and multi-resolution features make it a powerful tool for image fusion technology.This paper focuses on image fusion based on wavelet transform, making in-depthdiscussion and improvement on the existing algorithms from the following twoaspects, the low-frequency and high-frequency part of wavelet decomposition.Currently, there are two problems in image fusion based on wavelet transform, whichare the selections of the optima wavelet basis and wavelet decomposition level. Basedon these two points, the paper determines the optimal wavelet basis and the bestwavelet level by a large number of experiments. The paper also makes improvementon the rules of image fusion based on wavelet transformation. The low-frequency partuses the weighted average method. And the high-frequency part of the first uses theCanny operator to extract the edge of the high-frequency sub-image area, then makesthe fusion according to regional variance on the edge. The experimental results showthat, whether it is based on subjective visual evaluation or objective evaluation whichincludes information entropy, standard deviation, peak signal to noise ratio and spatialfrequency and so on, the proposed method is feasible. The main contents of this paper include the following aspects:(1) It introduces the background and significance of this paper, and then discussesthe image fusion including the simple basic concepts, development, application andresearch status.(2) It discusses the theoretical basis of image fusion, and describes the steps ofimage fusion as well as three levels of the current image fusion. Then studycomparatively of some common image fusion based on spatial and frequency domainmethods, points out the advantages and disadvantages of each method. Finally, thesubjective and objective quality evaluations of image fusion are briefly introduced.(3) It discusses the basic theoretical knowledge of wavelet transform, describesmulti-resolution analysis of wavelet transform and the image decomposition andreconstruction of wavelet transform in detail.(4) It studied the image fusion algorithm based on wavelet transform and theaffect of two aspects of image fusion to determine the optimal wavelet basis and thebest wavelet decomposition level. Then the image fusion algorithm based on Cannyoperator and the sub-regional variance is proposed and Matlab is employed tosimulation algorithm. The experimental results show that, whether it is based oninformation entropy, standard deviation and spatial frequency of objective evaluationcriteria, or subjective visual evaluation, the proposed selection methods are feasibleand effective.