超窄间隙焊接焊道截面的几何参数自动测量方法

Automatic Measurement Method of Geometric Parameters of Ultra-narrow Gap Welding Bead Cross-section

摘要

在超窄间隙焊接焊道截面的几何参数测量过程中,焊道截面轮廓数据特征点提取困难,而且人工测量几何参数效率低。针对该问题,提出一种超窄间隙焊接焊道截面轮廓拟合及几何参数自动测量的方法。首先,采用万物分割模型(segment anything model,SAM)对焊道截面图像进行分割,得到焊道截面边缘轮廓特征,再利用模板匹配算法自动识别分割图像;其次,对焊道边缘轮廓特征坐标点进行提取,并使用最小二乘法拟合焊道轮廓;最后,采用共面线性标定算法对相机进行标定,实现像素距离到实际距离的转换,并实现几何参数的定量检测。实验结果表明:该方法的曲线拟合决定系数R²的平均值大于94%,测量绝对误差均值为0.04~0.14 mm,相对误差均值为0.75%~2.77%,精度满足测量要求,效率相较于人工测量显著提高。

Abstract

When measuring the geometric parameters of ultra-narrow gap welding bead cross-section, extracting feature points from the welding bead profile data proves challenging and manual measurement of geometric parameters is inefficient. To address this issue, a method for fitting the cross-sectional profile of ultra-narrow gap welding beads and automatically measuring geometric parameters is proposed. First, the segment anything model (SAM) was used to segment the welding bead cross-section image, obtaining the edge contour features of the welding bead cross-section. Then, a template matching algorithm was used to automatically identify the segmented image. Next, the coordinate points of the welding bead edge contour features were extracted and the least squares method was applied to fit the welding bead contour. Finally, a coplanar linear calibration algorithm was used to calibrate the camera, converting pixel distances to actual distances to achieve quantitative detection of geometric parameters. Experimental results show that the average coefficient of determination R ² for curve fitting of this method exceeds 94%, the average absolute measurement error is between 0.04 mm and 0.14 mm, and the average relative error is between 0.75% and 2.77%. The accuracy meets the measurement requirements, and the efficiency is significantly improved compared to manual measurement.

关键词

超窄间隙焊接 / 焊道形貌 / 图像分割 / 边缘提取 / 自动测量 / 相机标定

Key words

ultra-narrow gap welding / welding bead topography / image segmentation / edge extraction / automatic measurements / camera calibration

参考文献

[1] 胡丹, 吕波, 王静静, 等 . 焊缝表面缺陷激光视觉传感HOG—SVM的检测方法[J]. 焊接学报, 2023, 44(1): 57-62, 70, 131—132.

[2] HU D, LÜ B, WANG J J, et al. Study on HOG—SVM detection method of weld surface defects using laser visual sensing[J]. Transactions of the China Welding Institution, 2023, 44(1): 57-62, 70, 131—132. (in Chinese)

[3] 滕彬, 范成磊, 徐锴, 等 . 厚板窄间隙焊接技术研究现状与应用进展[J]. 焊接学报, 2024, 45(1): 116—128, 136.

[4] TENG B, FAN C L, XU K, et al. Research status and application progress of narrow gap welding technology for thick plates[J]. Transactions of the China Welding Institution, 2024, 45(1): 116—128, 136. (in Chinese)

[5] 郑韶先 . 焊剂带约束电弧超窄间隙焊接方法研究[D]. 兰州: 兰州理工大学, 2007.

[6] ZHENG S X . The research on the technology of ultra—narrow gap welding with the flux strips constricting arc[D]. Lanzhou: Lanzhou University of Technology, 2007. (in Chinese)

[7] 郑韶先, 朱亮, 黄斌维, 等 . 焊剂带约束电弧用于超窄间隙的焊接[J]. 机械工程学报, 2009, 45(2): 219-223.

[8] ZHENG S X, ZHU L, HUANG B W, et al. Constricted arc by flux strips applied to ultra—narrow gap welding[J]. Journal of Mechanical Engineering, 2009, 45(2): 219-223. (in Chinese)

[9] NAGESH D S, DATTA G L . Genetic algorithm for optimization of welding variables for height to width ratio and application of ANN for prediction of bead geometry for TIG welding process[J]. Applied Soft Computing, 2010, 10(3): 897-907.

[10] 卓文波, 谭国笔, 陈秋任, 等 . 基于代理模型和NSGA—Ⅱ的超高强钢电阻点焊工艺参数多目标优化[J]. 焊接学报, 2024, 45(4): 20-25, 130.

[11] ZHUO W B, TAN G B, CHEN Q R, et al. Multi—objective optimization of resistance spot welding process parameters of ultra—high strength steel based on agent model and NSGA—Ⅱ[J]. Transactions of the China Welding Institution, 2024, 45(4): 20-25, 130. (in Chinese)

[12] 邹媛媛, 左克铸, 李鹏飞, 等 . 激光拼焊焊缝截面轮廓几何建模方法[J]. 激光与光电子学进展, 2017, 54(9): 228-233.

[13] ZOU Y Y, ZUO K Z, LI P F, et al. Geometric modeling methods for cross—sectional profile of welds in tailored blank laser welding[J]. Laser & Optoelectronics Progress, 2017, 54(9): 228-233. (in Chinese)

[14] HE W L, ZHANG A H, WANG P . Weld cross—section profile fitting and geometric dimension measurement method based on machine vision[J]. Applied Sciences, 2023, 13(7): 4455.

[15] 孙家豪, 张超勇, 吴剑钊, 等 . 基于神经网络的316L不锈钢激光焊焊缝形貌预测[J]. 焊接学报, 2021, 42(12): 40-47, 99.

[16] SUN J H, ZHANG C Y, WU J Z, et al. Prediction of weld profile of 316L stainless steel based on generalized regression neural network[J]. Transactions of the China Welding Institution, 2021, 42(12): 40-47, 99. (in Chinese)

[17] GAO Z M, SHAO X Y, JIANG P, et al. Parameters optimization of hybrid fiber laser—arc butt welding on 316L stainless steel using Kriging model and GA[J]. Optics & Laser Technology, 2016, 83: 153-162.

[18] 余杰, 蔡创, 谢佳, 等 . 基于响应面法的低碳贝氏体钢激光—电弧复合焊接工艺参数优化[J]. 中国激光, 2022, 49(16): 1602018.

[19] YU J, CAI C, XIE J, et al. Process parameter optimization for laser—arc hybrid welding of low—carbon bainite steel based on response surface methodology[J]. Chinese Journal of Lasers, 2022, 49(16): 1602018. (in Chinese)

[20] 张鹏贤, 李明, 冯毅 . 一种基于立体视觉的焊缝成形质量评价方法[J]. 上海交通大学学报, 2015, 49(1): 44—48, 55.

[21] ZHANG P X, LI M, FENG Y . An evaluation method of weld formation based on binocular stereo vision[J]. Journal of Shanghai Jiao Tong University, 2015, 49(1): 44—48, 55. (in Chinese)

[22] 何银水, 余卓骅, 李健, 等 . 基于视觉特征的厚板机器人焊接焊缝轮廓的有效提取[J]. 机械工程学报, 2019, 55(17): 56-60.

[23] HE Y S, YU Z H, LI J, et al. Effective weld seam profile extraction via visual features for robotic welding with thick steel plates[J]. Journal of Mechanical Engineering, 2019, 55(17): 56-60. (in Chinese)

[24] 吴曈, 胡浩基, 冯洋, 等 . 分割一切模型(SAM)在医学图像分割中的应用[J]. 中国激光, 2024, 51(21): 2107102.

[25] WU T, HU H J, FENG Y, et al. Application of segment anything model in medical image segmentation[J]. Chinese Journal of Lasers, 2024, 51(21): 2107102. (in Chinese)

[26] KIRILLOV A, MINTUN E, RAVI N, et al. Segment anything[EB/OL]. [2024—03—10]. https://arxiv.org/abs/2304.02643v1.

[27] 薛利军, 张虎, 李自田 . 基于削减搜索分支的快速模板匹配算法[J]. 计算机工程, 2003, 29(16): 87-89.

[28] XUE L J, ZHANG H, LI Z T . A fast algorithm for template matching based on search branches pruning[J]. Computer Engineering, 2003, 29(16): 87-89. (in Chinese)

[29] 张一梵 . 基于灰度相关的快速模板匹配算法研究[D]. 广州: 广州大学, 2022.

[30] ZHANG Y F . Research on fast template matching algorithm based on gray correlation[D]. Guangzhou: Guangzhou University, 2022. (in Chinese)

[31] 陈皓, 马彩文, 陈岳承, 等 . 基于灰度统计的快速模板匹配算法[J]. 光子学报, 2009, 38(6): 1586-1590.

[32] CHEN H, MA C W, CHEN Y C, et al. Fast template matching arithmetic based on image gray statistic[J]. Acta Photonica Sinica, 2009, 38(6): 1586-1590. (in Chinese)

[33] 雷冬, 杜文康, 朱国靖, 等 . 基于机器视觉方法的高铁桥梁监测技术研究[J]. 铁道工程学报, 2023, 40(3): 45-49.

[34] LEI D, DU W K, ZHU G J, et al. Research on the monitoring technology of high—speed railway bridge based on machine vision method[J]. Journal of Railway Engineering Society, 2023, 40(3): 45-49. (in Chinese)

[35] 杜春洋, 郭俊杰, 李昌非 . 基于改进Canny算子的焊缝图像边缘提取研究[J]. 激光杂志, 2023, 44(4): 74-79.

[36] DU C Y, GUO J J, LI C F . Research on edge extraction of weld image based on improved Canny operator[J]. Laser Journal, 2023, 44(4): 74-79. (in Chinese)

[37] 郝良耀, 王冬青, 许崇立 . 基于最小二乘多项式的焊缝图像拟合研究[J]. 电子设计工程, 2020, 28(23): 173-176, 182.

[38] HAO L Y, WANG D Q, XU C L . Weld extraction based on least square polynomial fitting method[J]. Electronic Design Engineering, 2020, 28(23): 173—176, 182. (in Chinese)

[39] 刘理想, 柏兴旺, 周祥曼, 等 . 电弧增材制造多层单道堆积的焊道轮廓模型函数[J]. 焊接学报, 2020, 41(6): 24-29, 36, 98.

[40] LIU L X, BAI X W, ZHOU X M, et al. Study on the weld profile model function of multi—layer single—pass deposition in wire and arc additive manufacturing[J]. Transactions of the China Welding Institution, 2020, 41(6): 24-29, 36, 98. (in Chinese)

[41] 尹英杰, 徐德, 张正涛, 等 . 基于单目视觉的平面测量[J]. 电子测量与仪器学报, 2013, 27(4): 347-352.

[42] YIN Y J, XU D, ZHANG Z T, et al. Plane measurement based on monocular vision[J]. Journal of Electronic Measurement and Instrument, 2013, 27(4): 347-352. (in Chinese)

基金资助

国家自然科学基金资助项目(62173170)