Research And Application Of Artifact Fragment Splicing Algorithm Based On Small Plane Area Feature

Relics and relics have far-reaching social significance as cultural relics that carry cultural,artistic and scientific research values.However,many relics have suffered different degrees of damage due to natural factors and human interference,making relic restoration and reconstruction crucial.Although traditional manual restoration methods suffer from long lead times,high costs and tendency to cause secondary damage,significant progress has been made in computer-aided virtual restoration technology with the development and application of 3D laser scanning technology.This advanced technology has not only greatly reduced the time cycle for restoration of cultural relics,but has also successfully avoided secondary human damage to fragile heritage fragments.Therefore,computer-assisted virtual restoration technology is of great significance and practical value in the field of cultural heritage conservation,providing new possibilities and directions for future restoration work.Digital fragment stitching is a key research area in computer-aided virtual restoration technology.In dealing with the problem of splicing non-thin-walled artifact fragments,the current main research method is to segment and extract features from the fracture surface,and then perform matching calculations.However,the surface of non-thin-walled artifacts that have been affected by weathering and human factors for a long time may be eroded and worn out,resulting in blurred details and insignificant geometric features on the fracture surface,making it difficult for the existing fragmentation algorithm based on fracture surface segmentation to effectively segment the surface of fragments with severe wear.To solve this problem,a splicing algorithm based on small planar area features is proposed.The algorithm segments the whole surface of the fragment by small planar area growth and uses the small planes as feature regions for matching,so as to improve the splicing accuracy and success rate.The main work of this paper is as follows:1 An algorithm for splicing artifact fragments based on small planar area features is proposed.The algorithm uses planar region growth to segment the entire fragment surface with small planes to avoid the problem that it is difficult to distinguish the fracture surface of worn fragments.Using the segmented small planes as feature regions,the small plane regions are characterized according to the dimensional features and anisotropic features to find the initial similar pairs of small plane regions.Finally,the coarse alignment of the fragments is performed by the 4PCS(4-Points Congruent Sets)algorithm,and then the exact collocation of the fragments is achieved according to the Iterative Closest Point algorithm ICP(Iterative Closest Point).The experimental results show that the small-plane feature description is simple and robust,which improves the success rate of wear fragment splicing and effectively reduces the penetration phenomenon between fragments.2 The constraint strategy of geometric constraint combined with RANSAC(Random Sampling Consistency)is proposed for the existence of one-to-many and two-to-two mismatched mis-matching correspondences in the found pairs of small planar feature regions.Most of the one-to-many mis-matches are eliminated by geometric constraints,and then the remaining mis-matching relations are eliminated by RANSAC.The experimental results show that the constraint strategy can better solve the mis-matching problem.