Super-resolution Method For Single Image

Super resolution technology is an important image research direction,which has a pivotal position in the field of image processing and imagequality enhancement. All along, the super resolution technique has steadydevelopment and importance widely used in the branch of medical imaging,image restoration, computer vision, pattern recognition, the public securityinvestigation, and so on.In view of the super-resolution algorithm for single image, the paperfirstly introduces the background, states the significance of thesuper-resolution technique, and summarizes the research status ofsuper-resolution, the related technologies and the main methods of study.After that, the paper gives the classification of the major super resolutionalgorithm, and tells the algorithm from three section which are based oninterpolation and reconstruction, on the study, and on the edge. Besides,the paper puts forward the corresponding algorithms or improvements onthe original algorithm and gives the results calculation of PSNR and thecomparison with typical methods, and shows the application. Finally, thepaper summarizes the whole work of innovations and raises the prospectfor the follow-up work.The paper proposes a method with a combination of convolutionsampling and anti-aliasing fusion based on the edge of the super-resolutionalgorithm for single image, through the analysis and integration of severalclassic algorithms. Considered in the method based on the reconstructionand on the study of super-resolution thought, innovative use of convolutionsampling technology and local self-similar properties generates twopictures in the edge and non-edge position with a high resolution image,and using the anti-aliasing analysis algorithm of graphics fusion technology, the paper makes a integration of two images to integrate theadvantages of position and result a high resolution image. Not only that, bychanging the sample order to get high frequency images and usinganti-aliasing technology which is more graphics direction related, the papermake the algorithm detail improvement, which is another innovation ofthis paper.