Ceramic Fragments Intelligent Reassembling Technology Based On Contour Features

Archaeological work is an important part of cultural tracing.Also,it is a powerful driving force to promote the in-depth development of historical research.There are tens of thousands of historical relics unearthed at the archaeological sites,and their appearances had been damaged to varying degrees.Reassembling and restoring broken cultural relics is a difficult problem in the industry.In this dissertation,computer technology is applied to the restoration and protection of cultural relics,in order to digitize the reassembling work of cultural relics.With taking the thin-wall cultural relics as the main research object,this dissertation studies the intelligent ceramic fragments reassembling method based on contour features from three aspects of preprocessing,contour processing,similarity matching and virtual reassembling.The main work is as follows:(1)Preprocessing method of three-dimensional cultural relic fragmentsThe data preprocessing is a key step in the intelligent restoration and protection of cultural relics.In order to transform the cultural relic fragments into three-dimensional data that is applicable to computers,this dissertation proposes a preprocessing method for 3D cultural relic fragments.First of all,this dissertation uses 3D laser scanning modeling method and photography modeling method to collect the point cloud data of the fragments.Secondly,the 3D data denoising method based on filtering and multi-head attention mechanism is used to remove the noise in the point cloud data of fragments obtained by digital acquisition.Finally,experiments are carried out on different data sets,and the results show that the preprocessing method of 3D cultural relic fragments is effective and applicable.In 3D point cloud data with relatively low noise level,the noise reduction effect and generalization ability of the method are improved.(2)Processing method of fragment’s point cloud contour based on neighborhood and space distanceFor thin-wall fragments,in order to effectively use their contour information,a method based on neighborhood for processing the contour of the fragment’s point cloud is proposed.Firstly,the method analyzes the neighborhood information of fragment’s point cloud data.And the recognition conditions are formulated based on the distribution characteristics of contour points in the object.Then the contour points are extracted from the fragment’s point cloud data by plane fitting and projection.Secondly,aiming at the disorder of point cloud data,the method of sorting points based on spatial distance is proposed.This method converts unordered contour points into orderly arrangement.Finally,in order to use the attributes and features of curve,the contour points of the fragment’s point cloud are fitted as differentiable functions,which realizes the transformation from contour points to contour curves.In this dissertation,the experimental results show that the method can produce effective contour data.(3)Iterative virtual reassembling method based on stepwise similarity matchingIn order to reassemble thin-wall fragments by using their contour features,an iterative virtual reassembling method based on stepwise similarity matching is designed and implemented.Firstly,the feature representation matrix of the fragment’s point cloud is constructed based on the feature points.Secondly,the similarity matching is divided into two steps: coarse comparison and fine comparison.In coarse comparison,the method calculates the difference of contour features by introducing Dynamic Time Warping(DTW)algorithm;In fine comparison,the method improves the Longest Common Subsequence(LCS)algorithm to calculate the similarity of contour segment feature sequences.Then,in the iterative transformation virtual reassembling method,the parameters of transformation are solved by Gaussian mixture distribution.And the results are continuously optimized by the idea of iterative transformation.Finally,this dissertation verifies the effectiveness of the method.Then the method is applied to the reassembling of ceramic relic fragments in order to provide a scheme for the archaeological work.