摘要
车载探地雷达技术通过发射和接收电磁波来探测和分析地下物体的特征,对保障地铁隧道的安全性起到重要作用;为了对地铁隧道缺陷进行精确检测,并提升检测的效率,构建基于Yolov5模型的车载探地雷达检测系统;首先采用零时校正、去直流、背景去除和图像增益方法对信号和图像进行去噪;然后基于Yolov5目标检测模型,引入特征网络金字塔和SPP-Bottleneck模块进行改进,并采用锚框对目标图像进行检测,最后构建基于Yolov5模型的车载探地雷达检测系统;结果显示,改进后的Yolov5模型在置信度相同的条件下,相较于原始模型具有更高的F_(1)值;在实例应用中,基于Yolov5模型的车载探地雷达检测系统F_(1)、精确度、召回率平均值分别为88.4%、87.3%和89.5%;该系统对缺陷进行检测的时间为0.3 s,相较于其他3种检测模型,效率分别提升了93.75%、84.2%和50.0%,更具有实际应用价值;此次研究解决了传统车载探地雷达技术存在的问题,对地铁的运营和维护具有重要的意义。
Vehicle mounted ground penetrating radar technology is used to detect and analyze the characteristics of subway tunnels by the emission and reception of electromagnetic waves,and it plays an important role in ensuring the safety and reliability of subway tunnels.In order to accurately detect the subway tunnel defects and improve the detection efficiency,an on-mounted radar detection system based on Yolov5 model is proposed.Firstly,the signal and image were first denoised through the methods of modifying the zero time,removing direct current(DC),background removal and image gain adjustment.Then,based on Yolov5 object detection model,the SPP-Bottleneck module and feature network pyramid are introduced to improve the performance of the system,the anchor box is adopted to detect the image of the target.The results show that the improved Yolov5 model has a higher F_(1)value compared to the original model under the same confidence level.In practical applications,the F_(1),accuracy,and recall average values of the vehicle mounted ground penetrating radar detection system based on the Yolov5 model are 88.4%,87.3%,and 89.5%,respectively.The Yolov5 object detection model has a detection time of 0.3 s,and its efficiency improves by 93.75%,84.2%,and 50.0%compared to the other three detection models,respectively,which has more practical application value.This study solves the problems existing in traditional vehicle mounted ground penetrating radar technology,and it is of great significance for the operation and maintenance of subways.
作者
曹志
高洪清
王威
刘华云
CAO Zhi;GAO Hongqing;WANG Wei;LIU Huayun(Nanjing Puzhen Rolling Stock Co.,Ltd.,CRRC,Nanjing 210000,China;School of Economics and Management,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;Nanjing Rail Transit Industry Development Co.,Ltd.,Nanjing 210000,China;Chengdu Tangyuan Electric Co.,Ltd.,Chengdu 610000,China)
出处
《计算机测量与控制》
2024年第4期61-66,共6页
Computer Measurement &Control