The Research On Terracotta Pottery3D Model Classification

A large number of Terra-cotta Warrior potteries were found during the excavation of Qin terracotta, these potteries have complex structures, large quantity and different shape of characteristics. Manual repair of them is time-consuming, low efficiency, and poor result. Using computer-aided method to sort, match and reassemble fragments is the mainstream method of repairing Terra-cotta Warrior pottery. In order to improve the Accuracy and Efficiency of Automatic classification, this paper proposes a three-dimensional model automatic classification method based on surface shape characteristics clustering. First, define the characteristics of the surface shape, then use Gauss curvature estimation method, quadratic surface fitting method to extract the characteristics of the surface shape, use K-medoid algorithm and EMD distance measure to cluster characteristic shape, found the center of clusters, build shape feature matching library of Terra-cotta Warrior pottery. Use C++language to achieve a classification system of Qin terracotta pottery. The main work of this paper is as follow:(1) Propose a new method used in existing Gauss curvature estimation method, quadratic surface fitting method to extract the cluster characteristic shape of pottery fragments.(2) Split15Terracotta model, Classify sub-model, extracte the characteristic shape of those sub-models to establish pottery fragments classification training set. Use K-medoid algorithm to cluster characteristic shape, found the center of clusters.(3) propose a three-dimensional model automatic classification method based on surface shape characteristics clustering. Use Gauss curvature estimation method, quadratic surface fitting method to extract the characteristics of the surface shape, use K-medoid algorithm and EMD distance measure to cluster characteristic shape, found the center of clusters, Experiments show that this method has a higher classification accuracy to terracotta pottery three-dimensional model of complex surface shape. (4)0n the basis of experiments, we can design the method to achieve the construction of pottery fragments classification system.