Traffic Police Gesture Command Recognition Based on Convolutional Neural Network AlexNet

Traffic police gesture signal recognition is of great significance in traffic management and urban intelligent transportation system, this paper researches the automatic recognition technology of traffic police command gesture signal based on the deep learning method, in the process of model training, the training set is used to train AlexNet, and the model is evaluated by the test set. The experimental results show that the model achieves high accuracy in the recognition of traffic police gesture signals. Finally, based on the trained model, a traffic police gesture signal recognition system is designed in MATLAB platform to realize the gesture signal recognition function on local images. The system can be used as a practical application tool for gesture signal recognition in traffic management, providing support for improving the efficiency and safety of urban traffic command.