| People always need to obtain the best quality images in many situations. But in the actual imaging process, the image is degraded by the restricts with imaging system, the external environment and the imaging technology and so on, and the images that we obtained are not certainly ideal. However, in the practical application, the quality of the image is required higher and higher. Therefore, how to restore the primitive image effectively or retreat the image that we expect becomes an important problem in image process.The super-resolution algorithm of image generally means a kind of image processing algorithm, it generates high quality image from a group of low quality or low-resolution images that belong to the identical scene. It means that as well as low frequency information restoration of the image, it restores the high frequency information above the cut-off frequency, and obtains more details and information to ensure the image from the restoration more similar the ideal image.The main contents of the study include such aspects as follows:Firstly, this article introduces the research situation about the super-resolution image restoration at home and abroad,and the study significance.Secondly, it introduces basic theory and main technologies about the super-resolution image restoration.Thirdly, it describes the varies methods about the super-resolution image restoration technology, emphasizes the analysis of the popular methods at present.Fourthly, it puts forward to an improved POCS algorithm by using wavelet technology to eliminate noises. Moreover, the experiments improve the advantages and necessarity of this method to restore the degraded image with low SNR.Finally, it introduces evaluating standard for the regular statistics method, and also proposes two evaluation criteria about edge-detection and Fourier Frequency spectrum comparison. Moreover the method in this thesis has been proved efficient by using these two new indexes. |
PDF Full Text Request