The Research And Application Of Video Image Super-resolution Reconstruction Algorithm

The process of acquiring high resolution image from one or more low resolution images and recovering technology of high frequency detail at the same time is called image super-resolution reconstruction. It can be applied to many fields, such as medical imaging, video monitoring and remote sensing. At the same time, multiframe super-resolution reconstruction technique can be used for video images.In the article, not only sparse representation is used for the single frame image reconstruction in the coupled feature space but neighbor embedding is used to reduce the time complexity of single frame image super-resolution reconstruction. And dynamic texture reconstruction method is used for multi- frame reconstruction.Firstly, in order to speed up the dictionary training, the algorithm of using the double dictionary in the coupled feature space for image super-resolution magnification is put forward. In order to reduce the complexity of the algorithm, there is a certain mapping relation between sparse representation coefficients for images proposed.Secondly, in order to reduce the time complexity of reconstructing and storage requirements, the algorithm based on sparse representation and neighbor embedding is proposed. To speed up the reconstruction, when calculating the distance between the dictionary atoms and image patches, the proposed algorithm is not each image patch to calculate. Instead image patches are clustered, and calculate the distance between dictionary atoms and the clustering center, then select image patches in the closest category. Calculated time of weight matrix can be reduced, and the efficiency of image reconstruction can be improved. Compared with some other algorithms, The resulting of PSNR is also improved obviously.At the end of the paper, dynamic texture synthesis algorithm is adopted for video image reconstruction. image segmentation algorithms are selected properly for region segmentation, and image super-resolution reconstruction is based on image region rather than the image patches. In order to reduce the time complexity, texture synthesis algorithm is only used part of the image frames for magnifing image, and then the bidirectional motion compensation algorithm is uesd for other image frames to complete the motion estimation. Finally, dynamic texture synthesis method is adopted to solve the problem of temporal coherence, and to obtain high resolution video image.