Development Of Urban Road Trafic Signal Detection And Navigation And Positioning System

With the increasing number of cars in our country,traffic safety and road pressure are becoming an urgent problem to be solved."Made in China made 2025" in the intelligent network of vehicles can be a good way to alleviate traffic safety problems.In order to make the intelligent vehicle safe to travel on the road,this paper has carried on the research to the intelligent vehicle traffic signal recognition technology and the navigation and positioning system according to the background of the urban complex road environment.The main research work and the result are as follows.First,the machine learning algorithm is used to detect the light box area of the intelligent traffic signal in front of the vehicle.First,we collect a large number of positive and negative samples of traffic lights in the actual road environment.Secondly,we unify the sample size,extract the HOG characteristics of the positive and negative samples,and train the support vector machines and cascade classifiers respectively.The supported support vector machines And the cascade classifier to detect the traffic lights of the road to realize the positioning of the light box.Through the experiment,the comparison between the two recognition performance shows that the HOG feature and the support vector machine can improve the overall efficiency of the traffic light box.93.5%,the average time-consuming 491ms,at the crossroads to meet the real-time requirements of vehicle travel;Finally,through the color statistical method to achieve the specific significance of traffic lights to identify.Secondly,by optimizing the K-means algorithm,the traffic signs of the front road are monitored and monitored,and the SIFT algorithm is used to realize the practical significance of traffic signs.Compared with other algorithms,this method is superior to the other algorithms.The average time is 120%and the average time is 120ms,which satisfies the real-time requirements of vehicle travel.Third,the development of intelligent vehicle positioning system.Secondly,based on the Baidu map API,Visual Studio environment,the use of MFC framework,the use of C/C++and JavaScript interactive programming,design,development,the use of serial communication,the use of serial communication,completed the data acquisition,analysis;The intelligent vehicle navigation electronic map;Finally,the completion of the integration of GPS/GIS processing to achieve the real-time positioning of the vehicle.Fourth,the intelligent vehicle control system software has been optimized to update the function.First,the traffic signal recognition,ban traffic sign recognition and map location function are added to the system.Then,the system function is integrated and optimized to improve the speed and stability of the system.The joint test between the modules is more reasonable,To provide the necessary conditions for the implementation of intelligent vehicle.