Zhong Qi: a Multi Sensor Image Fusion Algorithm Based on Multiresolution Analysis

Based on pixel level multi resolution analysis and current 3 multiresolution algorithms for Wavelet Transform (WT), Contour-let Transform (CT) and Non-Subsampled Contour-let Transform (NSCT), we present a new hybrid multi resolution algorithm. We use different types of images and extensive simulations combined with subjective and objective evaluation criteria to compare and analyze the pros and cons of each algorithm. We study: i) principles of multi-resolution analysis and the improvement of common multi-resolution analysis in WT, CT and NSCT algorithms and fusion rules; ii) the mixed framework of multiresolution analysis, and combine it with single multi resolution algorithm, and propose a hybrid multi resolution algorithm with two-dimensional multi-scale WT and NSCT; iii) multi focus image, infrared and visible images, medical images and images of four types based on WT, NSCT, WT and hybrid CT algorithm respectively, and determine the effectiveness of subjective analysis and objective parameters to validate the algorithm; and iv) we use Matlab platform to build a suitable demonstration and do comparative analysis of the image fusion system. KeywordsMultiresolution analysis; image fusion algorithm; multi-sensor data acquisition; wavelet transform; NSCT transform