The Research Of Digital Image Restoration Algorithm Based On Sample Block Filling

Digital image restoration is a technique that automatically restores lost or damaged areas in images or removes specific objects.In this digital and intelligent era,image restoration process can completely be dealt with by the computer.Digital image restoration has developed for 20 years,different types of image restoration methods emerge in endlessly,such as TV model,low rank matrix filling model and generate against network restoration model.Among several existing restoration algorithm models,this paper mainly introduces the restoration model based on sample block filling.This model has received more research attention due to its comprehensive consideration of image structure and texture information and good repair effect for large missing areas.At present,although there are many improved models in the part of priority calculation of the sample block to be filled and the part of confidence value update,the repair effect is not very good when the image texture is relatively rich.For this reason,this paper proposes a method to better judge whether the selected sample block is the best sample matching block by adding the direction gradient constraint to the pixels in the sample block in the process of finding the best sample matching block for the sample block to be repaired,so as to ensure that the texture and structure information of the known image area can be well spread and extended.The experimental results show that the improved algorithm in this paper can effectively guarantee a good visual effect when repairing images with abundant texture and structure,whether by using objective evaluation or PSNR evaluation index.During the repair process,we found that before the repair algorithm was implemented,the specific location of the repair area needed to be manually specified and the mask image needed to be made.Then the damaged area was repaired according to the accurate repair location information provided by the mask image.This process still required professional personnel to operate.Therefore,this paper based on the improved sample block filling repair,on the basis of combining text detection algorithm making mask information of repair area,the level of text automatic detection repair theory are introduced,analysis and discussion,and the text in the level of automatic positioning,removal,repair process simulation,the experimental results show that the automatic detection repair repair results of the algorithm diagram with the need for human to repair the area specified repair algorithm for visual effects,PSNR indexes also almost unanimously,proved effective repair automatic detection algorithm is feasible and practical.