一种基于卷积神经网络的受电弓滑板异常状态检测方法

打开文本图片集

中图分类号：U264.34 文献标志码：A 文章编号：2095-2945（2025）25-0131-05

Abstract：Inresponsetotheissuesoflowdetectionaccuracyandeffciencyintraditionalpantographslidercondition monitoringmethods，thispaperproposesapantographslideranomalydetectionmethodbasedonConvolutionalNeuralNetworks （CNNs）.Itcombinesdeepleaningtechniqueswithtraditionalimageprocessingtechnologies，extractsimagesofpantographders， andtheninputstheseimagesintoasliderconditiondetectionnetworktodeterminewhetherthecarbonslidersofthepantograph exhibitracksorabnormalwear.Ultimately，experimentsprovedtheefectivenessandaccuracyofthismethodinetecting abnormal conditions of pantograph carbon slide plates.

Keywords： Convolutional Neural Network（CNN）;pantograph； slide plate;condition detection; detection method

受电弓滑板的主要功能是与接触网导线直接接触并承担电流的采集任务。（剩余5248字）