Research On Traffic Object Recognition Method Based On Intelligent Terminal

Accurate collection of traffic object information is one of the basic conditions for traffic efficiency and traffic safety.Using the sensors of the intelligent terminal to collect road pedestrian status and vehicle information,to realize the identification and statistics of traffic participants,can greatly reduce the information collection cost of traffic objects.Use smart terminals to collect sensor data of road pedestrians and occupants,perform preprocessing,feature extraction,and classifier training on the data,establish road pedestrian and occupant state recognition models,and quickly identify different behavioral states.Then,a vehicle recognition model is established on the basis of driver behavior status recognition and statistics,to quickly identify the driving vehicle category in the traffic scene.The recognition results provide support for traffic efficiency and traffic safety related applications.The basic structures of the thesis are as follows:Firstly,The thesis analyzes the feasibility of traffic object recognition based on intelligent terminal,and designs a traffic object recognition architecture based on intelligent terminal.Sensor data is collected through an intelligent terminal,and filtering and simple averaging methods are used to perform gravity removal and reduce noise on the data,respectively.Then the time domain feature and the frequency domain feature extraction are performed on the data.The frequency domain feature extraction mainly uses discrete wavelet transform and fast Fourier transform.A feature selection method based on classification weights is designed to filter the features,and the classifier performs the filtered features.Training to establish a state recognition model of road pedestrians and occupants based on smart terminal sensor data to quickly identify different behavior states of people,including driving vehicles,road pedestrians standing,jogging and other states.Secondly,When the recognition result is a driving vehicle,perform cluster analysis on the motion characteristics of different types of vehicles,establish a vehicle recognition model,and quickly identify the types of driving vehicles in the traffic scene,including cars,bicycles,buses,etc.Finally,Based on the above research on road pedestrian status and driving vehicle recognition technology,a traffic object recognition system based on intelligent terminalsis designed and implemented.The system is used to identify traffic objects in real traffic scenes,and verify the accuracy of the classification algorithm in different scenes.Experiments show that the method designed in this thesis achieves an average recognition rate of 90.1% for the state of road pedestrians and driving vehicles,and the functions and performance meet the target requirements.