基于YOLO-V5的矿车桁架铆接孔定位技术研究

Research on riveting hole positioning technology of tramcar truss based on YOLO-V5

在矿车车厢自动化生产中,采用机器视觉技术来实现自动铆接工艺。为了对铆接孔进行准确识别与定位,采用深度学习中的YOLO-V5算法对铆接孔进行实时检测,结合最小二乘拟合圆法,在弥补了检测结果缺陷的同时,利用提取的铆接孔边缘坐标,实现了对铆接孔所在位置的准确定位。通过试验,证明了该方法的可行性与鲁棒性。通过对比试验,证明了算法的优越性。

In the automatic production of tramcar carriage, the machine vision technology was used to realize the automatic riveting technology.In order to accurately identify and locate the riveting holes, the YOLO-V5 algorithm in deep learning was used to detect the riveting holes in real time.Through the least square fitting circle method, the defects of the detection results were made up, and the edge coordinates of the riveting holes were extracted, the accurate location of riveting hole was realized.The feasibility and robustness of the method were proved by experiments, and the superiority of the algorithm was proved by comparative experiments.