Research On Reassembly Method Of Terracotta Warriors Fragments Based On Multi-feature

Cultural relics are the carriers of our country's long history and culture.However,influenced by natural disasters,man-made destruction and other factors,most of the ceramic cultural relics present a broken or incomplete state.The use of computer-assisted cultural relic virtual restoration technology has the advantages of high speed,convenience and accuracy,and it has gradually become a research hotspot in the field of computer graphics.Since most of the fragments discovered are of different shapes,with more and more broken types,it is difficult to design a reassembly method that can solve all types of fragment reassembly problems.The existing reassembly methods based on single geometry drive have the problems of single feature,complex descriptors,low matching accuracy and slow speed.Therefore,this thesis combines multiple features to study the problem of fragment recombination with different degrees of fault surface missing.The research work in this thesis mainly includes:(1)Aiming at the problem of reassembly of cultural relic fragments with relatively rich fracture surface information,a virtual reassembly method based on spanning tree cost and geometric constraints is proposed.Firstly,the n-features with significant concavity and convexity on the fracture surface of the fragments are extracted by the curvature function.Furthermore,the topological reconstruction is performed,the n-th order weighted undirected complete graph and its minimum and maximum spanning trees are constructed with the Euclidean distance between the spatial locations of the feature points as the weights,with the cost of spanning trees as the adjacency constraint,quickly screens the potential matching fragments;then,constructs the feature string with the main curvatures of the feature points,and then introduces the Hausdorff distance to measure the similarity between the two feature strings.it can effectively find the matching pieces;Finally,the quaternion method is used to align the fragments coarsely,and then the iterative closest point algorithm is used to achieveprecise alignment.The experimental results show that a good reorganization effect can be achieved for the fragment model of cultural relics with relatively complete fracture surface information.(2)In view of the problem of Reassembly of cultural relic fragments with certain defect of fracture surface leading to the loss of geometric information,A new method for Reassembly fragments based on the matching of characteristic points of fracture surfaces is proposed.Firstly,the improved ISS feature detection method is used to extract potential feature points of fragment fracture surfaces;secondly,the covariance matrix of geometric features of adjacent feature points is calculated to construct feature descriptors.Then,the logarithmic Euclidean Riemann method is used as the similarity measure criterion,the initial point pair set is obtained by the bidirectional nearest neighbor method,and the optimal matching set is obtained by eliminating mismatching pairs by CCA.Finally,the least square method is used to calculate the rigid body transformation matrix to align the fragments,and then the ICP algorithm is used to achieve precise alignment.The experimental results show that this method is simpler than the traditional method in descriptors and higher in matching accuracy,which can effectively improve the efficiency and accuracy of this kind of fragment reassembly.