Surface Attributes Driven Volume Segmentation

Currently,the mainstream fabrication techniques have difficulties in yielding models with diverse surface attributes or large size in one pass owing to their craft limitations.Aiming at these problem,an alternative is to partition 3D objects into single-attribute volumetric parts,which are printed with different materials separately,and finally assemble those parts to form the original model entity.This kind of work is called volume segmentation algorithm.Volume segmentation algorithm can not only expands the scope of application of 3D printing,but also improves the transportation efficiency of model entities.In this thesis,a novel volume segmentation algorithm is proposed.Our algorithm generates feasible model segmentation results from model input with predefined surface attributes,in which each volumetric part is associated with a single surface attribute.Our method uses radial basis function surface reconstruction algorithm and differential evolution algorithm to segment model and optimize the surface of parts,so that each part obtained by segmentation can be manufactured by 3D printing technology and assembled into the original model with minimal conflict.The main work of this paper is as follows:(1)We propose a tetrahedral mesh model segmentation algorithm guided by the implicit surface generated by the radial basis function surface reconstruction algorithm.The coordinates and normals of boundary vertices belonging to a certain region are used to generate the implicit surface,and this implicit surface is considered as dividing interface of current region to separate the tetrahedrons inside the model.(2)We propose a surface optimization algorithm based on the initial segmentation results obtained by implicit surface segmentation.By adding the offset surface constraint in the differential evolution algorithm process,interface of each part is applied to o generate a segmentation interface close to the minimal surface.(3)We propose a binary segmentation algorithm based on the output from surface optimization algorithm.Our approach computes a division plane and divide unextractable volumetric part into smaller sub-volume through the well-design plane.This step includes three parts: determining the testing sequence of volumetric parts,extractability examination of volumetric parts and binary segmentation for unextractable parts.This thesis provides experimental results for each step of our approach to prove the effectiveness and robustness of our volume segmentation algorithm,as well as its advantages over existing methods.