Research And Application Of Image Inpainting Technology In The Red Cultural Relics Of Gannan

Digital image inpainting is a hot research topic in machine vision,more and more researchers were involved in the study of image inpainting.It makes the image inpainting get a rapid development.However,because of the complexity of the damaged area,there is not a perfect algorithm can apply to all of the repair condition so far.To complete the adaptive image inpainting,the premise is able to image automatic segmentation of the damaged area.Now most of the damaged area is determined by manually.This paper studied the adaptive image segmentation by clustering technology and application of the theory of PDE in image inpainting based on the red cultural relics of gannan.summarized as follows:(1)K-Means Clustering is a typical unsupervised fuzzy clustering method,It has the advantages of simple principle,fast convergence speed,but over reliance on the initial cluster centers,and easy to fall into local value.The grey wolf optimizer algorithm is a new metaheuristic,has good ability of exploration and development.This paper proposes a hybrid clustering algorithm based on Grey Wolf Optimizer and K-means.The experiments show that the hybrid clustering algorithm has good global search capability and faster convergence speed,improves the stability and accuracy of K-means.(2)Adaptive extraction of damaged area image is a hot topic,this paper proposes the hybrid clustering algorithm is applied to image segmentation technology,realizes the adaptive image extraction of cultural relics damaged,improve the efficiency of image inpainting.(3)Research on image restoration technology based on partial differential equations,and using the partial model in the theory of differential equations on adaptive extraction of image to repair damaged cultural relics.