Fast Algoritihm For Two-dimensional Four-channel Non-separable Orthogonal Wavelet And Its Application In Image Fusion

Image fusion is an important subject under the background of information science research.At the same time it's a classification and data fusion in more mainstream,from the beginning in 70s,the research of image fusion research has continued.Image fusion is a comprehensive modern science and technology.The imaging principle and the particularity of the sensor's optical properties and other restrictions,a single sensor cannot obtain comprehensive information from the real scene,there must be some detailed information is not reflected,so the need from multiple sensors in many aspects of information extraction,which can reflect the real-time data from the comprehensive in.Multi source image can be divided into the following categories:time series image;multi-sensor image;multi focus image.Multi focus image is the same sensor zoom,the scene of a different bit of information capture,so as to obtain a number of clear and more detailed focus on the image information.The image data obtained from different sensors in all aspects can get rich information of the scene,it is helpful to understand the details of the image more comprehensive,the obtained from different sensor image data integrated,results contain comprehensive information,no single sensor method is given in this way,image fusion to remove the contradiction and redundancy of information may have multiple sensors to capture information,thus improving the accuracy of reliability,the new image of the image and get better image quality.In this paper,we present a fast algorithm for two-dimensional four channel non separable wavelets and give its mathematical proof.Firstly,the properties and characteristics of the two dimensional wavelet are studied in this paper.Secondly,the theoretical derivation and analysis of the fast algorithm show that the proposed algorithm is faster than the two dimensional non separable wavelet transform and two-dimensional fast Fourier transform.The decomposition of non-separable wavelet filter is constructed by group,and compared with the results of wavelet,non-separable wavelet transform to do image filtering can fully extract image information in each direction.Based on the fast algorithm of two-dimensional non-separable wavelet,this paper proposes a multi-focus image fusion scheme based on two-dimensional four channel non-separable wavelet.In this paper,a fusion strategy based on neighborhood correlation between pixels is adopted and applied to the image fusion of two-dimensional four channel non-separable wavelet.In the low frequency fusion,the strategy based on the regional variance that can preserve edges and details effectively.So that the synthetic image of the strategy will be clearer and save rich detail.This strategy can be effectively used for the multi-focus image fusion based on two-dimensional four channel non-separable wavelet.Comparing with DB2 wavelet and non-sampling non-separable wavelet from the actual fusion algorithm,the results of visual effects and varieties of image performance data,and through the analysis of experimental results,it can be concluded that the two-dimensional four channel non-separable wavelet fusion algorithm and its image fusion application can obtain the result image that is more clearly than the separable wavelet and non-sampling non-separable wavelet.It reflects good characteristics of fusion in preserving image edge details and data fusion index.