Vehicle In Front Of The Anti-collision Warning Methods

The car is one of the most important element in modem road traffic system,the widespread use of cars in promoting human society is fast and convenient,at the same time also brings many negative impacts,for instance,the possibility of traffic accidents in any time.Total mileage is more and more with the development of the society,kilometers long,more and more cars,the driver number also rapid growth,the traffic accident every year is also high.This paper mainly studies three aspects:to detect the vehicle ahead,to detect the people ahead,Measuring the distance between the obstacles ahead,finally,giving a warning when the car in the range of obstacles.This paper take two method of the shadow of motor vehicles and AdaBoost algorithm to detect the car ahead.The first method according to the highway gray mutation position locate the vehicle ahead.But the stability of algorithm remains to be improved.Then this paper tries to use AdaBoost algorithm to detect the vehicle ahead.Experiments show that the algorithm has better stability,and real-time performance.Based on the gradient direction histogram(HOG)combined with support vector machine(SVM)method to detecrt pedestrians ahead.After filling the camera parameter,Perspective projection model is used to calculate the local and the distance of the obstacles ahead.To warn the obstacles of the safe distance with secondary warning.The results show that,the development of Front collision warning system meet the system requirements in the rate of obstacle detection and the relative error range.largely provide the help in aspect of driving safety.