物探化探计算技术2026，Vol.48Issue(1)：29-35,7.DOI:10.12474/wthtjs.20241028-0002
基于PCA和欧式聚类的埋地排水管道声呐点云去噪技术
Point cloud denoising technology of buried drainage pipeline sonar based on PCA and European clustering
摘要
Abstract
With the progress of urbanization,the phenomenon of underground drainage pipes running with full water is increasing,and the traditional detection methods are difficult to adapt to the health assessment of pipes in the environment with full water.Sonar detection technology provides a new solution for health assessment of high-water-level and full-water-drainage pipes,but the generated point cloud data is often contaminated with noise.In this paper,a point cloud denoising method based on PCA and adaptive European clustering is proposed.First,PCA was used to identify and fit the principal components of the inner wall of the pipeline;then,noise was removed,and deposition lines were extracted using adaptive European-style clustering,which retained the pipeline's key structural features.Through field application of the Dongguan drainage pipeline,the results show that this method can effectively improve the quality of sonar point cloud data,accurately extract the pipe wall and sediment line,and meet engineering requirements.关键词
排水管道/声呐点云/PCA/自适应欧式聚类/去噪/沉积检测Key words
drainage pipeline/sonar point cloud/PCA/adaptive euclidean clustering/denoising/deposition detection分类
通用工业技术引用本文复制引用
WANG Hantao,CHEN Siyu,SUN Hongliang,SHANG Fangze,ZHOU Da..基于PCA和欧式聚类的埋地排水管道声呐点云去噪技术[J].物探化探计算技术,2026,48(1):29-35,7.基金项目
地质灾害防治与地质环境保护国家重点实验室自主研究课题(SKLGP2022Z011) （SKLGP2022Z011）
微分方程四川省自然科学基金创新研究群体(2023NSFSC1984) （2023NSFSC1984）
教育部"春晖"计划项目(2020703SC001) （2020703SC001）
