摘要
本文研究了基于深度学习与三维点云技术的接触网故障自动识别技术,通过激光扫描设备采集三维点云数据,结合多传感器融合技术进行精确预处理,利用卷积神经网络+循环神经网络(Convolutional Neural Networks+Recurrent Neural Network,CNN+RNN)深度学习模型,实现了高准确率和召回率的故障识别。案例显示,该技术在轨道交通中提高了故障诊断的准确性和效率,预计该方法将在智能化维护领域有更广泛的应用。
In this paper,we study the automatic fault identification technology of contact network based on deep learning and 3D point cloud technology,which collects 3D point cloud data through laser scanning equipment,combines with multi-sensor fusion technology for accurate pre-processing,and utilizes convolutional neural networks+recurrent neural network(CNN+RNN)deep learning model to achieve high accuracy and recall rate.The case shows that the technique improves the accuracy and efficiency of fault diagnosis in rail transportation,and it is expected that the method will have a wider application in the field of intelligent maintenance.
作者
李桥
LI Qiao(Xi’an Railway Vocational and Technical Institute,Xi’an Shaanxi 710026,China)
出处
《信息与电脑》
2024年第19期154-156,共3页
Information & Computer
基金
西安铁路职业技术学院2024年度立项课题《基于三维点云辅助分析的接触网图像故障诊断研究》(课题编号:XTZY24K09)。
关键词
深度学习
三维点云
接触网
故障识别
智能化维护
deep learning
3D point cloud
catenary
fault identification
intelligent maintenance
