摘要
针对机械式和光电式方法检测钢丝绳缠绕状态存在受外部环境影响大、传感器安装工艺复杂且要求高、加装传感器系统安全风险增加等问题,提出了基于人工智能与数字图像处理技术的自动在线检测方法。首先,采用BASNet网络识别包含轮毂下边缘的显著性区域;然后,通过轮廓跟踪算法和霍夫变换检测目标边缘和基准边缘;最后,通过计算目标边缘和基准边缘的位置关系并与已有限制参数对比,判断钢丝绳缠绕状态。实验结果表明:该方法正确率可达95.86%,处理单张图像耗时约0.60 s,初步具备工程应用条件。
Aiming at the problems of mechanical and photoelectric methods for detecting the winding state of steel wire ropes,such as the great influence of external environment,the complex and demanding sensor installation process,and the increased security risk of installing sensor system,an automatic online detection method was proposed based on artificial intelligence and digital image proces-sing technology.Firstly,the salient area containing the lower edge of the wire rope wound carrier was identified using BASNet network.Then,the target edge and the reference edge were detected by contour tracking algorithm and Hoff transform.Finally,the position relationship between the target edge and the reference edge was calculated and was compared with existing restricted parameters to judge the wire rope winding state.The experiments results show that the accuracy of the method is up to 95.86%,and the processing time of single image is about 0.60 s,which is preliminarily qualified for engineering application.
作者
郭威
徐兴华
崔小鹏
邱少华
欧阳斌
GUO Wei;XU Xing-hua;CUI Xiao-peng;QIU Shao-hua;OUYANG Bin(National Key Laboratory of Science and Technology on Vessel Integrated Power System, Naval Univ. of Engineering, Wuhan 430033, China)
出处
《海军工程大学学报》
CAS
北大核心
2022年第2期7-12,共6页
Journal of Naval University of Engineering
基金
国家自然科学基金资助项目(61701517)。
关键词
显著性检测
轮廓跟踪
霍夫变换
缠绕状态
saliency detection
contour tracking
Hough transform
winding state