This study addresses the issue of the existence of a considerable number of ancient ceramic fragments within the gene pool of ancient ceramics in Jingdezhen, as well as the limited efficacy of manual restoration techniques. To this end, an ancient ceramic restoration method based on the stitching of ancient ceramic textures through image processing is proposed. By employing the optimal single responsiveness matrix and state transfer equation, coupled with a random sampling strategy to ascertain the precise matching points and a consistency checking mechanism to preclude erroneous matching, the optimal single responsiveness matrix is devised to meticulously transform and stitch ancient ceramic textures. The dynamic programming idea is employed to identify the optimal stitching path, thereby enhancing the quality of the stitching and facilitating the precise, seamless, and natural integration of the ancient ceramic texture. The average peak signal-to-noise ratio of the stitched image of ancient ceramics is 58.7554, and the mean square error is 0.0866, which demonstrates the efficacy of image processing technology in the restoration of ancient ceramics and facilitates the intelligent advancement of cultural protection.
Read full abstract