附加增值条件的移动最小二乘法的点云孔洞修补  

Additional value-added conditional moving least squares method for point cloud hole repair

作　　者：兰猗令 康传利 王宁 杨佳乐 陈进启 Lan Yiling;Kang Chuanli;Wang Ning;Yang Jiale;Chen Jinqi(College of Geomatics and Geoinformation,Guilin University of Technology,Guilin 541006.China)

机构地区：[1]桂林理工大学测绘地理信息学院,桂林541006

出　　处：《红外与激光工程》2023年第2期406-415,共10页Infrared and Laser Engineering

基　　金：国家自然科学基金(41961063,42064002)。

摘　　要：由于扫描设备局限或模型结构复杂等因素导致点云模型出现孔洞,这严重影响模型的后续处理。针对点云孔洞的修补问题,文中提出了一种附加增值条件移动最小二乘法的点云孔洞修补方法。首先提取封闭的孔洞边界,通过密度分析进行迭代切片,不仅削弱点云分布不均的影响,还提高模型细节特征的保留程度;再将离散群点投影至拟合曲面,投影点集二次拟合以获取拟合面节点,保证有足够的边界邻域节点为基础进行孔洞修补;最后利用附加增值条件移动最小二乘法对孔洞进行迭代修补,并对增值点云进行曲率约束,从而达到契合原始模型空间特征的重建。实验采用人为在四个点云模型上制造不同类型的孔洞,并与现有的四种方法进行对比,验证所提方法的有效性,结果表明,文中方法相较于现有的四种方法,完整率、准确率提高了1.83%以上,配准均方根误差与平均曲率均方根降低了68%以上,对比证明了文中方法对于点云模型孔洞具有较强的适用性,可为重建三维点云模型提供可靠信息。Due to the limitations of scanning equipment or the complexity of the model structure, holes appear in the point cloud model, which seriously affects the subsequent processing of the model. To address the problem of point cloud hole repair, this paper proposes a point cloud hole repair method with an additional value-added conditional moving least squares method. Firstly, the closed hole boundary is extracted and iteratively sliced through density analysis, which not only reduces the impact of the uneven distribution of the point cloud, but also improves the retention standard of detailed features of the model;Besides, the discrete group of points is projected onto the fitted surface, and the projected point set is fitted twice to obtain the nodes of the fitted surface to ensure that there are enough boundary neighborhood nodes for hole repair;Finally, the holes are repaired by using the additional value-added conditional moving least squares method. Meanwhile, the curvature of the value-added point cloud is constrained, so as to achieve the reconstruction that fits the spatial characteristics of the original model. In the test, different types of holes are artificially made on four point cloud models, and the effectiveness of proposed method is verified by comparison with the existing four methods. The results show that, compared with the four existing methods, the completeness and accuracy of this method are improved by more than 1.83%,and the root mean square error of the alignment and the root mean square of the curvature are reduced by more than 68%, which proves the applicability of this method for point cloud model holes, which can provide reliable information for the reconstruction of 3D point cloud models.

关 键 词：点云模型 孔洞修补 增值条件 曲率约束 移动最小二乘法 迭代切片 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]