面向低成本便携扫描设备的三维建模与编辑技术

With the maturing scanning device hardware technology, a variety of low-cost portable scanning devices continue to emerge. However, the follow-up of 3D data processing technology is relatively slow, and so the current 3D modeling and editing techniques are often unable to deal with raw scan point clouds acquired by common people, severely limiting the development of portable scanning devices and various 3D applications. Under a huge demand for three-dimensional models in E-commerce, education, health care, food industry, household and entertainment etc, this project is dedicated to the research that includes digitization through low-cost portable scanning devices, generation and preprocessing of original raw scans, extraction of reconstruction priors, formation of priori constraints, detection of geometric characteristics, structural features, skeletons and their interrelations from objects represented by point clouds, and feature-driven automatic registration, modeling and editing. This project aims to break through the bottleneck of traditional 3D model creation technology, develop automatic three-dimensional modeling and editing techniques for highly non-ideal raw scans, and make significant progress on the theoretical research of understanding and expressing point clouds.

随着扫描设备硬件技术的日渐成熟，各种各样的低成本便携扫描设备不断涌现，然而后续的三维数据处理技术却相对滞后，现有的建模与编辑方法往往无法处理普通大众随意采集的粗糙原始点云，严重限制了便携扫描设备的大规模推广和各种3D应用。本项目以电子商务、教育、医疗、餐饮、家居和影视娱乐等对三维模型的巨大需求为导向，致力于通过低成本便携扫描设备将各种物理实体数字化，快速生成原始点云并进行先验提取和其约束构建；通过对点云所表示的物体几何特征、结构特征、骨架特征和物体间相互关联特征的分析，提取全局特征；实现特征驱动式的点云自动配准；突破传统的模型创建瓶颈，提出基于不理想点云的自动建模与编辑技术，在点云理解与表示的理论研究方面取得重要进展。

现阶段各种各样的低成本便携扫描设备不断涌现，但三维数据处理技术却相对滞后，现有的建模与编辑方法往往无法处理普通大众随意采集的粗糙原始点云，严重限制了便携扫描设备的大规模推广和各种3D 应用。本项目以电子商务、教育、医疗、餐饮、家居和影视娱乐等对三维模型的巨大需求为导向，通过低成本便携扫描设备将各种物理实体数字化，快速生成原始点云并进行先验提取和约束构建；通过对点云所表示的物体几何特征、结构特征、骨架特征和物体间相互关联特征的分析，提取全局特征；实现特征驱动式的点云自动配准；突破传统的模型创建瓶颈，提出基于不理想点云的三维自动建模与编辑技术，在点云理解与表示的理论研究方面取得了重要进展。