基于模糊推理的复杂形状物体不完整点云数据分类策略及模糊估值研究

At present, there exist incomplete point cloud data in the process of 3D reconstruction. The traditional repair technology of point cloud data carries out simple repair or single repair, which cannot meet the requirements of application. It has become a bottleneck problem for solving the incomplete point cloud data repair in 3D reconstruction technology..The general model of classification and fuzzy estimation for incomplete point cloud data of complex shape objects is proposed in this project, adopting the fuzzy inference system with characteristics of data driving, nonlinear solving and prior knowledge fusion. Starting with the study of data mining and fuzzy inference, based on incomplete point cloud data parameter characteristics of complex shape objects, the optimal classification fuzzy inference system model is constructed for the point cloud data. The classification strategy for incomplete point cloud data is determined by analyzing the relationship between each parameter of point cloud data. The fuzzy estimation model is determined, combining with prior knowledge by studying geometric features of complex shape objects and classification of point cloud. Point cloud classification and valuation calculation of fuzzy inference are realized by the program design. Incomplete denture theory of point cloud data is verified and its model is improved. The project is expected to give important theoretical development and practical application to improve estimation calculation precision and efficiency of incomplete point cloud data of complex shape objects and promote the development of oral repair medicine. ....

目前在三维重建过程中一直存在不完整点云数据的问题，传统的点云数据修补技术只能完成简单的修补或者修补某一个或某一类模型，无法满足应用的要求。如何解决不完整点云数据的修补问题成为三维重建技术亟待解决的瓶颈问题。.本项目提出具有数据驱动、解决非线性及融合先验知识等特点的模糊推理系统对复杂形状物体的不完整点云数据进行分类和模糊估值的通用模型。从数据挖掘和模糊推理研究入手，通过对复杂形状物体不完整点云数据参数特征的研究，建立用于点云数据的分类最优模糊推理系统模型；通过点云数据各参数之间的蕴含关系分析，确定不完整点云数据的分类策略；通过复杂形状物体几何特征与点云分类研究，结合先验知识确定模糊估值模型；通过程序设计实现模糊推理的点云分类和估值计算，进行不完整义齿点云数据的理论验证和模型改进。本项目对于提高复杂形状物体不完整点云估值计算的精度、效率，推动口腔修复医学的发展，具有理论意义和实际应用价值。

目前在三维重建过程中一直存在不完整点云数据的问题，传统的点云数据修补技术只能完成简单的修补或者修补某一个或某一类模型，无法满足应用的要求。如何解决不完整点云数据的修补问题成为三维重建技术亟待解决的瓶颈问题。. 本项目提出具有数据驱动、解决非线性及融合先验知识等特点的模糊推理系统对复杂形状物体的不完整点云数据进行分类和模糊估值的通用模型。从数据挖掘和模糊推理研究入手，通过对复杂形状物体不完整点云数据参数特征的研究，确定影响曲面拟合的关键参数；从不规则模型几何特征入手，对三维点云数据重建曲面过程中点云的密度、曲率、法向量进行了研究，研究了模糊推理系统中变量的作用域和隶属函数，建立用于点云数据的分类最优模糊推理系统模型；通过点云数据各参数之间的蕴含关系分析，研究了不完整点云数据所形成不同形状的孔洞特征，确定不完整点云数据的分类策略，将不完整点云数据形成的形状分成5类；通过复杂形状物体几何特征与点云分类研究，结合先验知识确定模糊估值模型，通过程序设计实现每类的估值计算，针对不同类型给出相应的修补方案，考虑增加数据点的数量和法向量约束调整。同时进行了不完整义齿点云数据的理论验证和模型改进。本项目提高了复杂形状物体不完整点云估值计算的精度、效率，推动口腔修复医学的发展，具有理论意义和实际应用价值。