Study Of Enhancement Technique Of Video Image Super Resolution

Enhancement technique of video image super resolution is the most economic and convenient method improving image resolution and it's optimistic applicational prospect has attracted wide concern. Enhancement of video super resolution amalgamates the information of several low resolution ratio video flow form equal dynamic scenes, and reconstructs a high spatial resolution video flow.Two different researching methods of enhancement technique of video image super resolution include: frequency domain method and spatial domain method. Frequence domain is mainly based on Spectral De-aliasing. Spatial domain includes interative backporjection method, POCS method, non-uniform spatial sample interpolation method, iterative backporjection method, set theory restoration method(POCS), statistic restoration method(MAP and ML), and blending MAP / POCS method, etc. Spatial domain treats the complicated motion model together with corresponding interpolation, interative and filtering re-acquisition. And its linear spatial domain observation model, which is comparatively near to the acturally acquired image, relates to entire and partial movement, optics blur, intra movement blur, spatial variable points spread function, non-ideal acquisition, and so on. Therefore, spatial domain method is more valuable in research and application.Wavelet theroy is developed rapidly in recent years, and wavelet is widely and successfully applied. This thesis foucses on super resolution image wavelet interpolation method. The improvement of wavelet super resolution, relying on orthonormal wavelet transform and the analysis of mulit-resolution, and under the frame of wavelet multiscaling decomposition and according to the known non-uniform interleaving hits, is the estimation of different scale space, then the estimated wavelet ratio is used as the interpolation for the unknown points on the super resolution grid.The simulation experiment based on MATLAB suggests that the wavelet interpolation method to enhance image super resolution is effective and efficient.