The invention discloses a track safety detection system based on improved AlexNet. An abnormal vibration signal judgment module of the track safety detection system determines that a driving vibration signal of each driving vibration signal detection area in each selected historical period is a normal driving vibration signal or an abnormal driving vibration signal; the vibration signal data set establishing module forms a model data set; the data enhancement module performs data enhancement on the model data set, and divides vibration signal data in the model data set after sample expansion into a training set, a verification set and a test set; the convolutional neural network model construction module constructs a convolutional neural network; a model training module uses the training set to train an AlexNet model; and the signal test module determines the track safety condition of the to-be-tested driving vibration signal test area. According to the invention, full-time and full-domain detection can be carried out on the track, and early warning can be carried out on external construction, track abnormity, personnel invasion and other problems.
本发明公开了一种基于改进AlexNet的轨道安全检测系统,它的异常振动信号判断模块确定每个行车振动信号测区各个选定历史时段的行车振动信号为正常行车振动信号或异常行车振动信号;振动信号数据集建立模块形成模型数据集;数据增强模块对模型数据集进行数据增强,并将样本扩充后的模型数据集内的振动信号数据划分为训练集、验证集和测试集;卷积神经网络模型构建模块构建卷积神经网络;模型训练模块使用所述训练集对AlexNet模型进行训练;信号测试模块确定待测行车振动信号测区的轨道安全状况。本发明能对轨道进行全时全域的检测,能对外界施工、轨道异常、人员入侵等问题做出预警。
Track safety detection system and method based on improved AlexNet
基于改进AlexNet的轨道安全检测系统及方法
2023-06-23
Patent
Electronic Resource
Chinese
IPC: | B61K Andere Hilfseinrichtungen für Eisenbahnen , OTHER AUXILIARY EQUIPMENT FOR RAILWAYS / B61L Leiten des Eisenbahnverkehrs , GUIDING RAILWAY TRAFFIC / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
ALEXNET-BASED INSULATOR SELF-EXPLOSION RECOGNITION METHOD
European Patent Office | 2021
|Research on Road Adhesion Condition Identification Based on an Improved ALexNet Model
DOAJ | 2021
|Convolutional Neural Network GNSS-R Sea Ice Detection Based on AlexNet Model
Springer Verlag | 2022
|