Research On Multi-source Image Fusion Algorithm Based On Multi Resolution Analysis

Multi-source image fusion refers to improvement of some images through research and improved fusion algorithm and fusion rule, from different types of image sensors capture images into a new image, a piece of information than any single image obtained by the image sensor to be more complete, rich and reliable. Currently, multi-source image fusion technology has spread to the social life of all walks of life, especially those with broad application prospects and important significance in the military affairs. Contents of the article focused on three levels of multi-source image fusion of pixel-level image fusion, and multi- resolution analysis is the main method of analysis at this level.By studying the pixel level in multi-resolution analysis theory and implementation, and current research hotter single multi-resolution wavelet transform algorithm, contourlet transform and nonsubsampled contourlet transform, and finally propose a new hybrid multi-resolution algorithm. Finally, simulation-based comparative test different types of images, a lot of combination of subjective and objective evaluation criteria to compare and analyze the pros and cons of each algorithm. The main research contents are as follows:A study of the principles of multi-resolution analysis and implementation, focusing on the common multi-resolution analysis of wavelet transform, contourlet transform and non-sampling algorithm and the fusion rules under contourlet transform improvements;In the framework of mixed multi-resolution analysis guidance, combined with the nature of a single multi-resolution algorithm, this paper proposes a hybrid multi-resolution algorithm, namely the two-dimensional multi-scale wavelet transform and nonsubsampled contourlet transform;Multi-focus image, remote sensing image, medical imaging and infrared and visible light images, sensing image based on four types of images were compared wavelet transform, NSCT transform, wavelet transform and contourlet hybrid algorithm in this article, given the subjective and objective analysis parameters verify the validity of the algorithm;Using Matlab platform, set up a convenient demonstration and comparative analysis of image fusion system.Through the above research, this thesis proposes a new hybrid multi-resolution algorithm--based on two-dimensional multi-scale wavelet transform and nonsubsampled contourlet transform, meanwhile, Improve handling of high and low frequency coefficient algorithm fusion rule. Finally, through a large number of experiments comparing simulation and analysis verify the effectiveness of the proposed algorithm.