The Design Of Preceding Vehicles Detection And Distance Measurement System Based On Machine Vision

With the rapid development of highway construction and increasing per cap ita consumption of cars in our country, the problem of traffic safety has become an unavoidable issue. Therefore, the ability of detection and distance measurement of preceding vehicle in real-time is an important measure of safe driving and autonomous navigation, and also a hot field of study in intelligent transportation systems.Although there are many classic vehicle detection algorithms, such as, Inter-frame difference method, Optical flow method, but they are not suitable for detecting preceding vehicles. This thesis propose an effective detection algorithm according to the characteristics of the preceding vehicle’s shadow, and select the method of distance measurement based on machine vision(monocular and binocular) by comparing the advantages and disadvantages of the method based on different sensors, achieve to measure the distance between the vehicle and the nearest vehicle, providing accurate assist driving information to the driver. This thesis mainly reasearch results are as follows:(1)Introduce the vehicle detection algorithm based on a variety of different ways, analysis of various algorithms application scenarios and the advantages and disvantages. Elaborate the limitation of radar and ultrasonic, highlighting the distance measurement based on machine vision has a wide application prospect.(2)This thesis presents an adaptive segmentation algorithm, based on the shadow of vehicles close to the bottom of vehicles stably, split the vehicle from the complex practical background, and generate hypothesis region of the vehicle. According to the symmerty of vehicle tail, calculate the symmetry measure by edge detection of sobel operator, in order to filter out the false vehicles that have been generated, reserved the real vehicle.(3)The imaging principle of camera is studied, using Zhang plane calibration method to calibrate the intrinsic parameters of the camera. Then use the distance mearsurement method of monocular visual and binocular parallax respectively to measure diatance of the preceding vehicles on real-time.(4)Based on Lab VIEW platform to develop the system software of preceding vehicles detection and distance measurement that based on machine vision. Using Lab VIEW vision library to complete the development of each module, and do the system experiments in the actual road, the results show that the system can detect preceding vehicles more accurately, and the nearest distance between the two vehicles can be measured within 10 meters, the farest distance can be measured about 50 meters, meet the safty requirements of 50 meters distance.