结合人工结构引导和图像块匹配的萌坑村清代壁画修复  

Research on Restoration of Qing Dynasty Architectural Murals in Mengkeng Village Based on Artificial Structure Guidance and Image Blocks

作　　者：蒋超 蒋振 潘敬婷 JIANG Chao;JIANG Zhen;PAN Jingting(College of Arts,Anhui University of Finance and Economics,233030,Bengbu,Anhui,China)

机构地区：[1]安徽财经大学艺术学院,安徽蚌埠233030

出　　处：《淮北师范大学学报（自然科学版）》2024年第3期59-65,共7页Journal of Huaibei Normal University：Natural Sciences

基　　金：安徽省哲学社会科学规划项目(AHSKQ2022D147)。

摘　　要：为萌坑村清代建筑壁画相关研究提供更为清晰的壁画图像信息,并为相似受损壁画的修复方法提供参考,提出一种融合人工引导和稀疏图像块模型算法的壁画数字化修复方法。从图像信息损失区域、损失程度与已知信息3个方面,对萌坑村清代建筑壁画的损伤现状与类型进行分析。针对壁画结构和内容缺失问题,通过加入人工结构引导的方式以明确修复区域边界。在人工结构引导下,基于图像块与稀疏表示模型算法实现损坏壁画区域的修复。萌坑村M-H-002壁画修复结果表明,所提出方法修复效果更为准确,视觉效果更佳,且在峰值信噪比和结构相似性方面,相较于传统的CDD(Curvature-Driven Diffusion)和Criminisi更优。In order to provide more explicit mural image information for research related to the Qing dynasty architectural murals in Mengkeng Village,and to supply a reference for the restoration method of similarly damaged murals,a mural digitization restoration method integrating manual structure guidance and image block sparse model algorithms is proposed and verified.Firstly,the damage status and types of Qing dynasty architectural murals in Mengkeng Village were analyzed in terms of image information loss area,loss degree and known information.Secondly,for the problem of mural structure and content loss,manual guidance was added to clarify the restoration boundary.Finally,based on the manual guidance,the damaged mural areas were restored by utilizing the image block and sparse representation model.The results of the mural restoration M-H-002 in Mengkeng Village show that the above-mentioned method is more accurate,and has better visual effects.It is superior to CDD and Criminisi in terms of peak signal-to-noise ratio and structural similarity.

关 键 词：人工结构引导 图像块匹配 清代建筑壁画 数字化修复 壁画修复 

分 类 号：K854.3[历史地理—考古学及博物馆学]