The Research on Recognition of Cultural Relics Material Information Based on Image

Abstract

In the process of realizing the virtualization of cultural relics, it is of certain significance to improve the research on the authenticity of cultural relics. and the way to realize it is to make the model of the virtualization of cultural relics match the perception of the corresponding real material. To solve this problem, an image-based classification algorithm for cultural relics material information is proposed, in which the material information should include the material of great majority of the cultural relics surface. In this process, the main purpose is to judge the metallicity and the reflectivity of the cultural relics surface. Through these surface material information, relevant programs can restore the real appearance of cultural relics in a better way. These classifications can be realized by depth learning. After random selection of pictures, the model can finally converge to about 95% of the material accuracy.