Research On Realistic 3D Modeling For Dense Point Cloud Of Cultural Relics

Chinese culture has a long history,leaving behind countless cultural treasures as the times change.With the passage of time and environmental changes,cultural relics are slowly dying out.Digitization of cultural relics can make them permanent,and the creation of a realistic 3D model of cultural relics is an important foundation for the digital reconstruction of cultural relics that have been damaged by the wear and tear of time.To fully restore the original appearance of cultural relics and obtain a digital model of cultural relics with a sense of reality,fine 3D modeling and realistic rendering are both essential.At present,in the research of 3D modeling of cultural relics,there are still many environmental noise points around the point cloud of cultural relics,the large amount of data of dense point cloud affects the reconstruction efficiency and the lack of texture information of the reconstructed model shows unrealistic,etc.These problems limit the development of digital conservation of cultural relics,and are also a problem that needs to be solved urgently.In order to further improve the digital conservation of cultural relics,this paper conducts a relevant study based on the above discussed problems,and the research work is as follows:(1)The 3D laser scanner,which can collect surface data without touching the object,is used to collect the model data of cultural relics,and for the complex and dense point cloud data containing environmental noise,the noise characteristics are analyzed,and the random sampling consistency algorithm and radius filtering algorithm are used to process the point cloud data of cultural relics in turn.After this denoising process,the final point cloud data of cultural relics without noise points is obtained,which lays a good data foundation for the subsequent streamlining and reconstruction.(2)Aiming at the situation that there is a lot of redundant data in the point cloud of dense cultural relics,the point cloud reduction algorithm is studied.In the process of researching the simplification algorithm,according to the characteristics of the balanced distribution of the left and right subtrees of the K-Dimensional tree(K-D tree),the randomness of the selection of the cluster center of the traditional K-Means clustering algorithm is improved.Aiming at the time-consuming problem of K-D tree backtracking query,the spatial grid method is combined with the K-D tree search algorithm.The grid method is used to segment the original point cloud data,and then the K-D tree is constructed for the cultural relic model data.Experiments show that the algorithm can effectively improve the search efficiency of k-neighborhood of point cloud data.(3)Aiming at the problem that the clustering reduction algorithm cannot identify the feature points,a point cloud reduction algorithm that retains features is proposed.The information entropy is introduced into the calculation of the estimated normal vector angle of the point cloud data,and the feature points are identified and retained through the information entropy value.Experiments show that this algorithm greatly reduces the amount of point cloud data and improves the reconstruction efficiency while maintaining the accuracy.(4)The greedy projection triangulation reconstruction algorithm is used to reconstruct the surface of the cultural relic point cloud model.In view of the fact that the reconstructed 3D model of the cultural relic lacks texture information and is not realistic,the OpenGL platform is used for texture mapping and the Blinn-Phong lighting model is introduced for photorealistic rendering..Experiments show that after the illumination model is introduced,the texture of the cultural relic model is clear,the detail features are obvious,and it is close to the real performance.