计算机工程与应用 ›› 2022, Vol. 58 ›› Issue (16): 249-256.DOI: 10.3778/j.issn.1002-8331.2101-0127

• 图形图像处理 • 上一篇    下一篇

基于机器视觉的钢管壁厚在线检测方法研究

涂德浴,刘坤,朱庆,刘庆运   

  1. 1.安徽工业大学 机械工程学院,安徽 马鞍山 243002
    2.特种重载机器人安徽省重点实验室,安徽 马鞍山 243032
  • 出版日期:2022-08-15 发布日期:2022-08-15

Research on Online Detection for Wall Thickness of Steel Pipe Based on Machine Vision

TU Deyu, LIU Kun, ZHU Qing, LIU Qingyun   

  1. 1.College of Mechanical Engineering, Anhui University of Technology, Ma’anshan, Anhui 243002, China
    2.Anhui Province Key Laboratory of Special Heavy Load Robot, Ma’anshan, Anhui 243032, China
  • Online:2022-08-15 Published:2022-08-15

摘要: 在钢管生产过程中,需要对钢管的壁厚进行实时测量,以检验所生产的钢管是否符合规格。针对人工测量钢管壁厚中所出现的测量效率低、易产生疲劳且无法实时测量等问题,基于机器视觉测量技术,提出一种钢管壁厚在线检测方法。该方法首先采集钢管断面图像,然后对采集到的钢管图像进行预处理,使用Canny边缘检测算子检测钢管断面内外圈边缘特征,最后使用改进后的随机霍夫圆检测算法检测断面轮廓,根据设计的钢管壁厚测量方案计算得到各处钢管壁厚。改进的随机霍夫圆检测由于采用分区采样以及筛选出与边缘轮廓契合度最高的候选圆的方式,提高了原检测算法的检测精度与效率。经过实验验证,该方法测量精度高,效率高,能够满足对钢管壁厚在线检测的要求。

关键词: 机器视觉, 图像预处理, 改进随机霍夫变换, 圆检测算法, 钢管壁厚测量

Abstract: In the process of steel pipe production, the wall thickness of steel pipe needs to be measured in real time to check whether the produced steel pipe meets the steel pipe specifications. Aiming at the problems of low measurement efficiency, easy to fatigue and inability to achieve real-time measurement in manual measurement of wall thickness of steel pipe. Based on machine vision measurement technology, an online detection method of wall thickness of steel pipe is proposed. Firstly, collect the image of steel pipe section, then performs pretreatment operations on the collected steel pipe image. The Canny edge detection is used to detect the edge features of the inner and outer circle of the steel pipe. Finally, the improved random Hough transform circle detection algorithm is used to detect the profile of the section. According to the designed measurement scheme of wall thickness of steel pipe, the wall thickness of steel pipe is got everywhere. The improved random Hough circle detection improves the detection accuracy and efficiency of the original detection algorithm, due to the use of partition sampling and screening of candidate circles with the highest degree of conformity with the edge contour. Experiments show that this method has high measurement accuracy and high efficiency, and can meet the requirements for online detection of wall thickness of steel pipe.

Key words: machine vision, image pretreatment, improved random Hough transform, circle detection algorithm, measurement on wall thickness of steel pipe