The invention provides a training method for a train detection model in an axle counting system, which can be applied to the technical field of deep learning, and comprises the following steps: obtaining a training data set which comprises multiple groups of training data and each group of training data carries a corresponding real label, and then carrying out iterative training on the train detection model in the axle counting system to obtain a train detection model in the axle counting system; and inputting the training data set into the train detection model in each iteration process to obtain a corresponding prediction label, and adjusting parameters of the train detection model according to the prediction label corresponding to the training data and the real label. According to the invention, the neural network model is introduced into the axle counting system, and then the neural network model is trained through the training data set, so that the state of the sensor and whether a train passes through are judged through the trained neural network model. Therefore, the efficiency of judging whether the sensor is abnormal or not and whether a train passes through or not can be improved, manual participation is not needed in the whole process, manpower resources can be saved, and the cost is reduced.
本申请提供一种计轴系统中列车检测模型的训练方法,可应用于深度学习技术领域,包括:获取训练数据集,其中,训练数据集包括多组训练数据,每组训练数据携带对应的真实标签,然后对计轴系统中的列车检测模型进行迭代训练,在每次迭代过程中将训练数据集输入至列车检测模型,得到对应的预测标签,根据训练数据对应的预测标签以及真实标签调节所述列车检测模型的参数。本申请在计轴系统中引入神经网络模型,然后通过训练数据集对神经网络模型进行训练,以通过训练好的神经网络模型对传感器状态以及是否有列车经过进行判断,从而可以提高判断传感器是否异常以及是否有列车通过的效率且由于整个过程无需人工参与,因此还可以节约人力资源,降低成本。
Training method and device for train detection model in axle counting system
一种计轴系统中列车检测模型的训练方法以及装置
2024-08-30
Patent
Electronic Resource
Chinese
Train low-speed axle counting detection system and method
European Patent Office | 2021
|