Super-Resolution Image Reconstruction Based On Non-Aliasing Contourlet

Super-resolution image reconstruction is a technique that can enhance an originalimage’s resolution with some certain algorithm. The process of Super-resolution imagereconstruction don’t excessively dependent on the hardware device, so it can reducethe costs which caused by the hardware upgrades. Super-resolution imagereconstruction contains two aspects: single image reconstruction and Multi-frameimage reconstruction. Both of them have been widely used in remote sensing, medicalimage processing, video surveillance systems, etc. A new algorithm ofSuper-resolution image reconstruction based on image registration, image fusion andContourlet was given in this paper.First of all, some popular algorithms of Super-resolution image reconstructionwere introduced. In order to improve the quality of the reconstructed image, thealgorithm was designed to improve the accuracy of image registration and reduce thefrequency aliasing of the Contourlet.Second, an algorithm of image registration based on PSO/Powell was given. Itcan meet the requirements of Super-resolution image reconstruction which needshigher accuracy of image registration. Then, in order to reduce the impact of frequencyaliasing, a new algorithm was designed based on non-aliasing Contourlet after weanalyze the cause of frequency aliasing. First, an anti-aliasing pyramid is used by thenew algorithm instead of the Laplacian pyramid. Then, the bi-cubic interpolation isused to do similar transformation in the high frequency domains, according to thesimilarity among the high frequency domains with different scales and differentdirections. Then, a higher-resolution image is obtained by inverse transform.At last, we did some experiments in Matlab software to verify the feasibility ofthe new algorithm. We also compared the result with the results of other similaralgorithms. The experimental results show the algorithm can get a better imagereconstruction effect.