Research On Image Super-resolution Restoration Algorithm Based On The Characteristics Of Redundancy Patches

The main task of super-resolution (SR) restoration is to obtain a high resolution imagefrom the given low resolution image. Super-resolution has been widely used in computer visionand many other areas. Obtaining high resolution image through better hardware is quite costly.Furthermore, the high-end camera installation is not feasible in many practical applications.Therefore, in practical application, it is desirable to increase the resolution of images fromexisting hardware device with an image super resolution technology, which is very economic,practical and effective.Obtaining the sequence of the same scene is very difficult in many practical applications,and sometimes it’s even impossible, so the research on single image is very meaningful. On thebasis of the basic algorithm research, this paper proposed an new algorithm of super resolutionusing the redundant patches which combined the classical SR constraints with based-examplesconstraints.The main contents of this paper are as follows:(1)We provide an introduction to background, development, current research, basicprinciple, classical super-resolution algorithm, and related performance evaluation of thesuper-resolution restoration.(2)The principle and the method of single image super-resolution restoration with sparserepresentation are introduced, followed with corresponding simulation experiment and analysis.(3)Without any external database, an improved algorithm of a single imagesuper-resolution is proposed, in which the characteristics of redundant patches is applied totraditional super-resolution reconstruction algorithm and classic SR and example-basedconstraints are combined. Experiment shows that the redundancy patches method can obtainhigher resolution image than the traditional methods.(4)We make a summarization and discuss the future directions of the applicationdevelopment about the single image super-resolution algorithm.