HIGHWAY TRAFFIC FLOW STATE RECOGNITION METHOD BASED ON DEEP NEURAL NETWORK

摘要

A highway traffic flow state recognition method based on a deep neural network, which relates to the technical field of intelligent traffic. The method comprises: classifying and defining a traffic flow state, carrying out noise reduction processing and feature extraction on an audio signal, carrying out modeling by means of a deep neural network (DNN) to obtain a deep neural network model for recognizing a highway traffic flow state, and pre-training the deep neural network model; then, tuning parameters of the deep neural network model; decoding a highway traffic flow state recognition model by means of a hidden Markov model (HMM); and finally, estimating an observation probability of the audio signal of different highway traffic flow states by means of the deep neural network model, and giving a recognition result of the highway traffic flow state according to the calculated probability. By means of the method, the problems of poor image analysis accuracy, a large amount of calculation for dynamic image analysis, etc. of monitoring traffic information using existing image analysis technology can be effectively solved.