Single Image Reconstruction Based On The Theory Of The Compressed Sensing

The image super-resolution reconstruction technology adopts the method of digital signal processing, can effectively reduce costs in all aspects, has great commercial and practical value. The current superresolution reconstruction technique has been widely used in security monitoring, medical imaging, radar imaging, pattern recognition and geological exploration and other fields.Compressive Sensing theory proposed for the study of image reconstruction technology provides a new way of thinking, greatly ease the digital signal data sampling, data transmission and storage process by the enormous pressure. This article focuses on a single image reconstruction method to explore under the theoretical framework.This article begins with a brief introduction to the image superresolution reconstruction technology and its development status, introduced the model of the algorithm. Then according to the Compressive Sensing and Sparse Representation theory, this paper studied the methods of single image super-resolution reconstruction based on sparse representation and joint dictionary, and proposed an improved method named Quxiangsi to create the set of sample digital images, and based on the Quxiangsi algorithm we studied the structure and behoof of classification dictionaries, and simulated the programs to test the feasibility and the effect. Finally, we compared the effect of our methods with the original one and find out that by using our method we can effectively reduce the cost of training the joint dictionary, and the dictionary we got from our method can help us to improve the quality of the image reconstructed by the original algorithm. By our way to train classification dictionaries bases on Quxiangsi algorithm we improved the efficiency of dictionary training and shortening the time cost under the premise to ensure the quality of the reconstructed image.