Research And Implementation Of Damaged Image Repair Technology

With the rapid development of digital image correlation technology,and more and more widely in the field of the application of digital image,digital image restoration technology is also increasingly active,for some old photos of crease and stain removal,post-processing of film and television works,photographic works of tailings rejecting has a variety of domains such as more practical application.At the same time,due to the use of computer repair processing,the image repair process can be repeated on the basis of maintaining the original appearance,until the most satisfactory repair results.Therefore,the study of image restoration technology has a very important practical significance.This thesis firstly analyzes the shape structure features of the image itself,pixel domain method,gray co-occurrence matrix method,texture spectrum method,autoregressive texture model method and other texture features,and then describes the extraction methods of various texture structure features.Repair BSCB model,TV repair model and CDD repair model such as the texture structure of image restoration algorithms,and based on image decomposition and the texture of the structure of the image restoration algorithm based on the sample is analyzed,based on the texture in the process of image restoration spread,and the problem of lack of texture priority based on the brightness of the image restoration algorithm is proposed,through the image gradient sample piece of the known information content,structure and intensity information,the combination of image block priority judgment,by means of adaptive template matching complete restoration of the missing image.The end of the paper via text scratches lack,lack of the lack of a single grain structure and complex texture structure lack of three groups of different types of images,respectively to Criminisi algorithm,feature matching algorithm,brightness priority priority algorithm for image restoration experiment simulation,repair the result image PSNR and SSIM values by calculation,complete the image inpainting results of data analysis.The experimental data show that the luminance priority algorithm has a good repair result after adding luminance factor to the priority calculation formula,although it increases a small amount of calculation.