三维时序点云模型的表面平滑与特征增强技术研究

项目名称： 三维时序点云模型的表面平滑与特征增强技术研究

项目编号： No.61300125

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 廖斌

作者单位： 湖北大学

项目金额： 22万元

中文摘要： 三维时序点云模型主要是通过高速三维扫描设备采集运动变形物体表面获取的。由于受扫描设备物理特性、扫描环境、光照等因素的影响, 得到的随时间连续演化的采样数据不可避免地会受到各种噪声的污染。需要对原始采集数据利用表面平滑与特征增强等方法进行处理得到普适性更好的三维时序点云模型。原始采集数据固有的采样缺失、各帧之间无时空对应、数据量大等诸多问题，给处理带来了巨大的挑战。本项目的研究内容包括三维时序点云模型的保特征快速表面平滑、三维时序点云模型的快速多层特征分割、三维时序点云模型的多尺度特征增强等技术。这些都是获取高质量三维时序点云模型的关键研究领域，也是三维时序点云模型编辑造型等几何处理的重要基础研究领域。本项目为以上研究内容提供了可行的研究方案，初步构建一个强壮的三维时序点云模型处理工具。

中文关键词： 三维时序点云；表面平滑；特征增强；噪声；特征分割

英文摘要： 3D time-dependent point cloud sequences are built using surface information of moving and deformation objects which is obtained by 3D high-speed scanning devices. Because of physical characteristics of scanning devices, environment of measurement and lighting, the sampling data with temporal evolution includes different types of noisy points. The raw sampling data should be processing using surface smoothing and feature enhancement, in the hope of improving the quality of 3D time-dependent point cloud sequences. It is a challenging computational problem for processing raw sampling data, since it contains missing data, lack of correspondence across frames and huge data amount. The research contents of the project are feature-preserving fast surface smoothing of 3D time-dependent point cloud sequences, fast hierarchical feature segmentation of 3D time-dependent point cloud sequences, and multi-scale feature enhancement of 3D time-dependent point cloud sequences. The research contents are key research aspects in improving the quality of 3D time-dependent point cloud sequences, and important fundamental research areas in the geometry editting and modeling processing of 3D time-dependent point cloud sequences. The feasibility study approaches for above research contents are also proposed and integrated into a robu

英文关键词： 3D time-dependent point cloud ；surface smoothing；feature enhancement；noise；feature segmentation