3D Surface Splicing Based on Principal Component Feature Extraction

In order to achieve automatic splicing for complex 3D surfaces, the parametric surface representation method for fractured surfaces of fragmented objects with thickness was analyzed and then a surface splicing method based on eigenvector of mixed subscript was proposed in this paper. In this method, geometric features of a surface were extracted based on principal components and directed subscripts and undirected subscripts were adopted to form mixed eigenvectors, and then the matching relationship between two surfaces was determined based on two similarity judgments. This method effectively lowered surface splicing error. Surface splicing experiments of broken cultural relic fragments verified the feasibility and effectiveness of this method.