Research And Implementation Of Cultural Relic Fragment Matching Technology Based On Multi-feature Of Fracture Surface

The study of cultural heritage is of great significance to our whole country and even the world.With the continuous development of information technology,with the help of computer-aided digital technology,the virtual splicing of cultural relics is becoming the mainstream.Compared with the traditional manual splicing method,digital virtual splicing is not only more efficient,but also can avoid the "intimate" contact with cultural relics,resulting in secondary damage.The research of digital virtual splicing of cultural relics is of great value in archaeology,medicine,artificial intelligence and other fields.For the problem of inaccurate feature extraction caused by local fragment missing and fracture surface damage,this paper proposes two splicing methods of cultural relic fragments,and develops a virtual splicing system of cultural relic fragments to verify the algorithm:(1)In order to solve the problem that the feature extraction of fracture surface is not accurate due to the lack of local fragments and the damage of geometric features in the process of Cultural Relic Fragment mosaic,a method of Cultural Relic Fragment mosaic based on speed up robust feature(surf)feature descriptor and Jaccard distance is proposed by this paper.Firstly,the 3D Canny operator is used to extract the edge contour of the fragment and the decorative lines on the surface of the model,and then the feature points of the fracture surface are extracted by constructing a multi-scale space.Secondly,in view of the high redundancy of feature descriptors and the high delay of Euclidean distance calculation,this paper proposes to construct surf feature descriptors with low redundancy,and then use Jaccard distance to compare the similarity of feature points to determine the optimal adjacency of fragments.Finally,the rotation parameters and translation parameters are calculated by iterative close point(ICP)algorithm to achieve more accurate matching of fragments.The experimental results show that,compared with the existing methods,the proposed method can effectively reduce the large stitching gap,penetration and other phenomena caused by the damage of the fragment model,and the stitching error is smal,so the fragment stitching can be realized efficiently.(2)For the broken fragments with thick fracture surface,in order to avoid the distortion and loss of feature points in the process of feature extraction,a method of Cultural Relic Fragment mosaic based on cost value edge and geometric constraints is proposed by this paper.Firstly,the fragments are triangulated,and then the cost value edge and small root heap are fused to simplify the mesh.Secondly,the local curvature threshold method is used to extract the feature points of the fracture surface,and the fast point feature histograms(fpfh)feature descriptor which can reflect the neighborhood information is constructed by differentiating the geometric parameters of the neighboring feature points.Then the Manhattan distance is defined as the constraint rule to realize the rough matching of similarity,and the geometric constraint based on the normal angle is used to realize the fine matching of similarity.Finally,the improved iterative closure point(ICP)algorithm is used to solve the variable parameters and realize the splicing of adjacent fragments.Experiments show that the proposed method has the characteristics of high stability,low redundancy and high timeliness while ensuring more accurate splicing.(3)A virtual splicing system of cultural relic fragments is desgined and developed by this paper.The test results show that the system can realize the splicing of adjacent fragments more accurately,which makes the work of cultural relic virtual restoration more efficient,and promotes and contributes to the digital virtual restoration technology of cultural relic.