基于机器视觉的石油管具螺纹缺陷检测系统  

Internal thread defect detection system of petroleum tubing based on multi-vision

作　　者：张耕培 窦萧寒 李刚[1] ZHANG Gengpei;DOU Xiaohan;LI Gang(School of Electronic Information and Electrical Engineering,Yangtze University,Jingzhou 434023,Hubei)

机构地区：[1]长江大学电子信息与电气工程学院,湖北荆州434023

出　　处：《长江大学学报（自然科学版）》2024年第6期120-126,共7页Journal of Yangtze University(Natural Science Edition)

基　　金：国家自然科学基金项目“随钻方位声波测井图像复原方法研究”(51604038)。

摘　　要：石油工业中的管具螺纹检测面临狭小空间、光线不足和几何复杂性等挑战,传统的人工检测方法效率低且错误率较高。为此,提出了一种基于机器视觉的自动化内螺纹检测方案,旨在提升检测精度与效率。通过优化图像采集系统,显著提高了图像采集速度和质量,确保在光线条件不佳时获取清晰的螺纹图像。采用圆柱模型拼接技术,将多个螺纹图像合成为全视野视图,增强了检测的全面性。在图像处理方面,系统结合YOLOv8深度学习模型,能够快速准确地定位和分类螺纹缺陷。研究结果表明:该技术在狭小空间内表现出色,提供了高效、可靠的解决方案,满足了石油管具行业对高效、精准检测的需求,显著提升了管具的安全性和可靠性,具有广泛的应用前景。Pipe thread detection in the petroleum industry faces challenges such as narrow space,insufficient light and geometric complexity.Traditional manual detection methods have low efficiency and high error rate.Therefore,this paper proposes an automatic internal thread detection scheme based on machine vision,aiming to improve the detection accuracy and efficiency.By optimizing the image acquisition system,the image acquisition speed and quality are significantly improved,ensuring the acquisition of clear thread images when the light conditions are poor.At the same time,the cylindrical model stitching technology is used to synthesize multiple thread images into a full-field view,which enhances the comprehensiveness of the detection.In terms of image processing,the system combined with YOLOv8 deep learning model can quickly and accurately locate and classify thread defects.The research results show that the technology performs well in a narrow space,provides an efficient and reliable solution,meets the needs of the oil pipe tool industry for efficient and accurate detection,significantly improves the safety and reliability of pipe tools,and has a wide range of application prospects.

关 键 词：石油管具 内螺纹 缺陷检测 机器视觉 

分 类 号：TE973.6[石油与天然气工程—石油机械设备] TP216[自动化与计算机技术—检测技术与自动化装置]