PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
February 26, 2026Optics and Precision Engineering0 citationsOpen Access

基于改进边界点排斥的激光点云模型孔洞修复

View Full Paper
LQLan QiupingMQMa QinghuaMHMEI Hong

Key Points

  • The aim is to develop an effective method for identifying and repairing holes in laser point clouds.
  • Utilized weighted local optimal projection algorithm to optimize point cloud distribution.
  • Employed normal vector analysis and local density constraints for precise hole boundary identification.
  • Implemented dynamic repulsion field model to generate filling paths based on topological relations.
  • Applied median filtering to ensure smoothness consistency between repaired and original point clouds.
  • Achieved a hole repair rate of 98.7% in the Camel model.
  • Point cloud density difference was only 4.2%.
  • Repair time was significantly faster than methods like LS-SVM, Poisson, PU-Net, and SeedFormer, being only 11.7% of SeedFormer's time.

Abstract

三维激光扫描技术虽能高效构建精确的三维模型,但在测量具有内凹曲面或狭窄缝隙的物体时,部分区域因光线遮挡而形成点云孔洞,影响模型的完整性与后续应用。传统修复方法多依赖插值或曲面拟合,易导致修复区域曲率不均、细节丢失。为解决该问题,提出一种基于改进边界排斥的激光点云孔洞识别与修复方法。采用加权局部最优投影算法优化点云分布,并通过法向量分析与局部密度约束精准识别孔洞边界,利用最大角度筛选排除伪边界点,避免将大曲率部位误判为缺损区域;同时,设计动态排斥力场模型,结合单圈邻接点拓扑关系生成填充路径,确保修复点云与原始模型的几何一致性;最后,利用中值滤波实现修复孔洞点云与整体点云平滑度的一致性。实验结果表明,本方法在多类复杂模型上均能有效识别与修复孔洞,其中Camel模型的修复率达98.7%,点云密度差异仅为4.2%,修复时间显著优于LS-SVM,Poisson,PU-Net及SeedFormer等方法,仅为SeedFormer耗时的11.7%。本方法在几何结构保持与细节修复方面具有明显优势,为激光点云孔洞修复提供了一种高效、可靠的解决方案。

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Qiuping et al. (2025) studied this question.

synapsesocial.com/papers/699fe37b95ddcd3a253e75cdhttps://doi.org/10.37188/ope.20253322.3475
Ask AI
Helpful
Bookmark
Share
View Full Paper