结合点云距离和角度双阈值的 桥梁拉索表面缺陷精确检测

打开文本图片集

关键词：桥梁拉索；表面缺陷检测；点云；感兴趣区域提取；法向量中图分类号：TP391文献标志码：ADOI：10.7652/xjtuxb202506016 文章编号：0253-987X（2025）06-0155-12

Precise Detection of Surface Defects in Bridge Cables Using Combined Point Cloud Distance and Dual-Threshold Angle Analysis

XIAXiaohua，ZHUYingshuo，QIU Fabo，CHENJian （Key Laboratoryof RoadConstruction Technologyand Equipmentof MOE，Chang'an University，Xi'an71oo64，China）

Abstract： To address the issues of mis-segmentation and high computational complexity in the application of point cloud segmentation algorithms to bridge cable surface defect detection，a precise detection method for bridge cable surface defects based on dual thresholds of point cloud distance and angle is proposed in this paper. First， a distance threshold was established between the cable surface points and the fitted cylindrical surface to facilitate preliminary defect detection. Then，a region of interest （ROI） extraction method for cable surface defects was designed， combining point cloud clustering and box filtering. Based on the preliminary detection results， the ROI was segmented using an improved Euclidean clustering algorithm coupled with box filtering. Finally，a defect boundary detection method was constructed based on the angle between the point cloud normal vector and the radial vector of the fitted cylinder. The defect points within the ROI were segmented by setting an angle threshold，enabling precise detection of the defect boundaries on the cable surface. Research results demonstrated that，compared with methods that do not perform ROI extraction or those that extract ROI without point cloud clustering，the proposed method reduced the point cloud data volume by more than 40% ，thereby shortening the detection time by over 15% . The average precision and recall rates for cable surface defect detection reached 96.83% and 94.42% ，respectively，with a F -score of 95.58% ， outperforming the comparative random sample consensus algorithm and the region growing algorithm. This validates the effectiveness and advancement of the proposed method.

Keywords： bridge cable；surface defect detection； point cloud; region of interest extraction； normalvector

拉索作为索承体系桥梁的关键受力构件，其承重和传递荷载能力是桥梁安全运行的关键[1]。（剩余18881字）