| Tree dimensional geometric matching and similarity research aims to achieve recognition, similarity measurement, and matching of three dimensional objects, by means of object analysis, transformation, and feature extraction. As a major part of this research, three dimensional surface matching is of great importance in a variety ares, such as registration of multiple range data sets, virtual reality, medical image registration and molecules structure design of drug.The basic problem of surface matching is to determine whether two 2d geometric objects (which are always surfaces of 3d geometric objects) are equivalent or not, Surface matching is an important task in computational geometry, and it plays an important role in robot and automatic control fileds.The majority of this dissertation focus on the research of three dimensional free-form surface matching and its related problems.In this paper it is given a brief survey of existing techniques of surface matching, including algorithms of feature-base surface matching method, model-based surface matching method and entire points-based surface matching method. Researches in related topics are also included, such as the presentation of free-form surfaces, the hierarchical presentation of surfaces, and techniques for feature detection.Besides the above mentioned work, a hierarchical feature-based surface matching method is proposed. The algorithm determines rigid transformation between the two given triangulated surfaces. We use points who's curvatures are larger than a threshold as the feature points. First, this surface matching method extracts features hierarchically, and then finds the correspondence between feature points, and computing the rigid transformation between the two surfaces to realize the surface matching. With the strategy of hierarchical matching, the speed of surface matching is highly increased. Experiments show that the algorithm is feasible, and performs fairly well in three dimensional surface matching. |
PDF Full Text Request