A Priori-based Adaptive Repair Method for Mural Images

Dunhuang Mogao Grottoes murals have a huge scale and a long history. Due to the joint influence of natural and human factors, murals have suffered serious damage. The murals have suffered from diseases such as armoring, falling off, smoking, cracking, fading and discoloration. How to repair these murals is a key problem. Aiming at the above-mentioned mural diseases, this chapter focuses on the mural noise removal, mural missing content completion and other aspects of repair research, and proposes a priori-based adaptive mural image restoration algorithm, which combines local total variation and variational regularization based on non-local graph to solve the ill-posed problems in image processing field. The algorithm is iteratively updated by the current image calculation. The experimental results show that our method achieves the same quality as the existing methods, but it has better robustness on Dunhuang murals which are complex in noise type.