A Study Of Virtual Restoration Techniques Of Damaged Ancient Tomb Murals

Ancient tomb murals are an important part of Chinese traditional mural paintings.The factors have been destroyed by natural and man-made factors for thousands of years.Therefore,cracks,fall-offs,and discoloration are often present when unearthed.In order to restore the original appearance,at this stage,it is mainly done manually and manually after painting.It not only depends on the individual experience of the technicians,but also takes time and effort.The computer-assisted restoration of murals in ancient tombs can provide references for artificial restoration,thereby increasing the efficiency and reducing secondary damage.Therefore,it is particularly important.Therefore,this article takes the ancient tomb mural images collected by Shanxi Museum as the research object,and conducts research according to the classification of disease types.The main tasks are as follows:(1)Aiming at the cracks in the complex forms of ancient tomb murals,a virtual repair method based on the combination of sample block repair algorithm and improved BSCB(Bertalmio-Sapiro-Caselles-Bellester,BSCB)model was proposed.Firstly,after the histogram is enhanced,cracks are acquired using the largest between-class variance method(Otsu);secondly,the connected-domain labeling algorithm is used in combination with opening and closing operations to separate cracks;and then an improved BSCB model is used to repair all cracks;The block image repair algorithm eliminates the traces of diffusion repair;finally,it uses color migration to recover the foreground color and obtain the final repair result of the crack.(2)Aiming at the shedding of complex forms in ancient tomb murals,a repair method based on a combination of structural continuous blocks and multi-scale transformation was proposed.Firstly,after the histogram is enhanced,the threshold segmentation is used to obtain shedding.In combination with the connected domain marker and the closed operation,the pixels in the shedding area are continuous,and a small amount of human-computer interaction is used to obtain the shedding area and the shedding boundary is determined;secondly,the confidence level is decreased and the confidence is matched.The items,data items,and gradient items determine the priority of the boundary point;again,the global search using the pixel difference and structural difference between the block to be repaired and the matching block results in the best matching block to replace the block to be repaired;then,the new complement point confidence item is updated.,re-defining the defect boundary until it is empty,get the rough repair result,and convert it to HSV color space;Finally,the non-subsampled contourlet transform is used to decompose the V component,and the low-frequency information is repaired by the suppression data item method.With its inverse transformation of high-frequency information,the results are synthesized with the H and S components and converted to the RGB color space,resulting in shedding final repair results.(3)Designed and implemented a virtual restoration system for the destruction of ancient tombs.The repair system adopts Graphical User Interface design and development in Matlab,and is divided into two modules: virtual repair and virtual restoration.Crack virtual restoration module mainly includes crack extraction,improved BSCB model repair cracks,sample block repair algorithm to eliminate the repair traces and foreground color recovery functions;shedding virtual repair module includes a combination of a small number of human-machine interaction extraction shedding,improved Cruminisi algorithm rough repair,HSV conversion,data item repair and RGB conversion functions.