| 29 | 0 | 62 |
| 下载次数 | 被引频次 | 阅读次数 |
针对石化设备点蚀量化评估难题，提出一种融合点云局部几何特征与双阶段聚类算法的点蚀检测方法。基于激光结构光扫描技术获取设备表面高精度点云数据，通过对点云数据进行噪声滤除、数据降采样等预处理后提取各点法向量、曲率等局部几何特征，并筛选出对点蚀敏感的特征组合。在检测流程中，首先采用K-Means聚类实现点蚀区域与非点蚀区域的粗分割，继而通过快速欧式聚类完成单个点蚀单元的精准定位与深度计算；通过二次聚类提取点蚀边缘轮廓，基于凸包算法与椭圆拟合实现开口面积的精确量化。实验结果表明，使用该方法测量304/316不锈钢点蚀深度的平均绝对误差为0.019 mm,最大误差控制在0.047 mm以内，可同步输出点蚀密度、开口尺寸、深度分布等多维参数，有效解决了传统人工测量效率低、二维图像缺失深度信息等问题。
Abstract：To address the challenge of quantitative evaluation of pitting corrosion in petrochemical equipment, a pitting corrosion detection method integrating the local geometric features of point clouds and a two-stage clustering algorithm is proposed. High-precision point cloud data of the equipment surface is obtained using laser structured light scanning technology. After preprocessing the point cloud data through noise filtering and data downsampling, local geometric features such as the normal vector and curvature of each point are extracted, and screen out the combinations that are scensitive to pitting corrosion. The detection workflow first employs K-Means clustering for coarse segmentation of pitting and non-pitting regions, followed by rapid Euclidean clustering to precisely locate individual pitting units and calculate their depths. The pitting edge contour is extracted through secondary clustering, and the precise quantification of the opening area is achieved based on the convex hull algorithm and ellipse fitting. The experimental results showed that the average absolute error for measuring the pitting depth of 304/316 stainless steel using this method was 0.019 mm, and the maximum error was controlled within 0.047 mm. It can synchronously output multi-dimensional parameters, including pitting density, opening size, and depth distribution, effectively solving the problems of low efficiency in traditional manual measurement and the lack of depth information in two-dimensional images.
[1] WANG Y,CHENG G.Quantitative evaluation of pit sizes for high strength steel:Electrochemical noise,3-D measurement,and image-recognition-based statistical analysis[J].Materials & Design,2016,94:176-185.
[2] ASTM G46—21 Standard Guide for Examination and Evaluation of Pitting Corrosion[S].
[3] GB/T 35013—2018 承压设备合于使用评价[S].
[4] WANG Y,CHENG G.Application of gradient-basedhough transform to the detection of corrosion pits in optical images[J].Applied Surface Science,2016,366:9-18.
[5] ZHAO C,ZHU H,WANG X.Steel plate surface defect recognition method based on depth information[C]//2019 IEEE 8th Data Driven Control and Learning Systems Conference (DDCLS).New York:IEEE,2019:322-327.
[6] 朱秀敏，黄磊.基于三维激光点云的零件表面缺陷检测[J].仪表技术与传感器，2022(7):56-60.
[7] 伍麟，郝鸿宇，宋友.基于计算机视觉的工业金属表面缺陷检测综述[J].自动化学报，2024,50(7):1261-1283.
[8] GUMILAR I,FAROHI F,MUNARDA M,et al.The combined use of terrestrial laser scanner and handheld 3D scanner for 3D modeling of piping instrumentation at oil and gas company[J].Journal of Engineering and Technological Sciences,2022,54(6):220603.
[9] 赵文慈，闫岩，汪红兵，等.基于点云处理的产品表面质量缺陷识别研究综述[J].冶金自动化，2022,46(5):20-35+102.
[10] DIMITRIOU N,LEONTARIS L,VAFEIADIS T,et al.A deep learning framework for simulation and defect prediction applied in microelectronics[J].Simulation Modelling Practice and Theory,2020,100:102063.
[11] DU J,YAN H,CHANG T,et al.A tensor voting-based surface anomaly classification approach by using 3D point cloud data[J].Journal of Manufacturing Science and Engineering,2022,144(5):051005.
[12] YAN K,LIU G,LI Q,et al.Corrosion characteristics and evaluation of galvanized high-strength steel wire for bridge cables based on 3D laser scanning and image recognition[J].Construction and Building Materials,2024,422:135845.
[13] LI R,JIN M,PAQUIT V.Geometrical defect detection for additive manufacturing with machine learning models[J].Materials & Design,2021,206:109726.
[14] WEINMANN M,JUTZI B,MALLET C.Feature relevance assessment for the semantic interpretation of 3D point cloud data[J].ISPRS Annals of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2013,2:313-318.
[15] MONNIER F,VALLET B,SOHEILIAN B.Trees detection from laser point clouds acquired in dense urban areas by a mobile mapping system[J].ISPRS Annals of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2012,I-3:245-250.
[16] 杨现辉，王惠南.ICP算法在3D点云配准中的应用研究[J].计算机仿真，2010,27(8):235-238.
[17] 孙吉贵，刘杰，赵连宇.聚类算法研究[J].软件学报，2008(1):48-61.
[18] 杨俊闯，赵超.K-Means聚类算法研究综述[J].计算机工程与应用，2019,55(23):7-14+63.
[19] CAO Y,WANG Y,XUE Y,et al.FEC:Fast euclidean clustering for point cloud segmentation[J].Drones,2022,6(11):325.
[20] GB/T 17897—2016合金和金属的腐蚀不锈钢三氯化铁点腐蚀试验方法[S].
[21] 刘书南，陈西江，花向红，等.邻域特征参数融合的点云轮廓特征点提取[J].遥感学报，2025,29(3):808-822.
基本信息:
中图分类号:TE65;TQ050.9
引用信息:
[1]申志远,杨锋,刘媛双,等.融合点云局部几何特征与双阶段聚类的点蚀检测方法[J].安全、健康和环境,2025,25(10):19-27.
基金信息:
国家重点研发计划(2022YFC3004502),氯碱化工生产装备隐蔽性损伤检测技术及装备研发; 中国石油化工股份公司科技部项目(325012),非平面点蚀结构光三维检测技术及腐蚀穿孔预测方法研究