Based On The Fracture Surface Of The Broken Matching Rigid Rehabilitation Research

Reassembling3D broken solids is a prominent and difficult problem in the fields of computer graphics, computer vision and pattern recognition. It can be applied in many fields such as archaeology, industrial design, clinic medicine and so on. More researches address fragments that have no thickness based on contour curve matching than fragments with thickness. We focuse on the reassembling broken solids based on surface matching, including fracture surfaces segmention, matching and fragments reassembly, all algorithms are verified in Terra Cotta Warriors and other fragments. Mainly works are as following:(1)We present a fracture surfaces segmentation algorithms based on normal vectors. First we segment the fragments into a set of surfaces bounded by sharp curves according to the angle of normal vectors of adjacent triangles.Then use perturbation value to distinguish the original surfaces and fractures. The algorithm is suited for fragments with sharp boundary.(2) We present a surface segmentation and contour curve extraction algorithm based on multi-scale curvature. We give a new method which first segments surface then spreads their initial boundary point and uses boundary tracking to get the final boundary. The algorithm is effective for fragments with sharp and complex boundary.(3) A fracture surfaces matching algorithm based on contour curve matching is presented. First we find pairs of matching surfaces based on their contour curve similarity and then we delete outliers according to normal vector continuous constraint and surface matching error. The algorithm is suited for surfaces matching when they have similar boundary.(4) A fracture surfaces matching algorithm based on multi-scale curvedness and3D histogram is presented. We give a new method which first divides fracture surfaces into several small regions whose centres are significant feature point and then uses3D histogram to compare region similarity. After that an exhaustive search method constrained by triangle similarity and voting scheme is used to match fracture surfaces.(5) A fracture surfaces matching algorithm based on integral invariants and feature region is presented. First fracture surfaces are divided into several featured region according to the concave or convex. Then we use principal component analysis (PCA) to obtain similar region pairs and eliminate outliers according to the distance and principal direction constraints. Finally we use ICP algorithm to verify the reliability of matching result.(6) We employ a sub-graph merging algorithm based on backtracking to merge all matched fragments until the object is reassembled.When there is a fragment which is not fusion, then backtracking.