Noisy disordered point cloud automatic registration method

• Main Authors: WANG CONGRAN, ZHAO YANGYANG, LEI JINBO, ZHAO XUAN, SUN SONG, WANG YONG, YANG FENG
• Format: Patent
• Language: Chinese; English
• Published: 01.05.2020
• Subjects: CALCULATING; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a noisy disordered point cloud automatic registration method. The invention relates to the field of point cloud data model construction. The method aims at the characteristics that the actual noisy point cloud data has multiple discrete point types and different characteristics; according to the method, a step-by-step filtering or smoothing method is adopted, so that the final point cloud registration efficiency and precision are greatly improved, the tedious operation of setting mark points by special mechanisms or people is not needed, multi-angle point cloud automaticregistration is realized, and the model reconstruction requirement of an actual scanning object is effectively met. 本发明提供了一种带噪无序点云自动配准方法，涉及点云数据模型构建领域，针对实际带噪点云数据具有离散点种类多和特征各异的特点，本发明采用分步滤除或平滑的方法，极大地提高了最终的点云配准效率和精度，不需特殊机构或人为设置标记点的繁琐操作，实现多角度点云自动配准，有效满足实际扫描对象的模型重建需求。