Study On Preceding Vehicle Detection And Distance Measurement Based On Computer Vision

China's economy keeps steady growth and the number of cars is increasing.In order to improve road safety and reduce traffic accidents,advanced driving assistance system(ADAS)came into being.The ADAS system can identify and track both static and dynamic objects in the process of vehicle driving,so as to warn the driver to be aware of the potential danger,and increase the comfort and safety of driving effectively.The preceding vehicle detection and distance measurement are two of the most important components in the advanced driver assistance system.Aiming at automatic preceding vehicle detection and intelligent collision avoidance,this paper studies on vehicle detection technology and vehicle range estimation method based on computer vision,through the analysis and comparison of various algorithms at home and abroad.In the aspect of preceding vehicle detection,a preceding vehicle detection method based on Haar-like feature and AdaBoost algorithm combined with Gaussian Mixture Model(GMM)is proposed in this paper.This method uses GMM algorithm for background modeling,filter out part of background,avoid exhaustive scanning of the entire image,eliminate the influence of target detection in complex background interference,can reduce the number of scanning window,so as to reduce the scanning time and increase the detection efficiency.The AdaBoost algorithm is used to select the Haar-like feature to generate the cascade classifier for vehicle detection,and verify the detected target,and finally get the accurate vehicle information.In terms of ranging,by calibrating the inside and outside parameters of the camera after we have detected the vehicle,and use the principle of visual projection model and geometric distance measurement to estimate the distance of the preceding car,so as to predict the possible collision events.ADAS as an important part of active safety technology,it could effectively improve the driving safety and the road safety,reduce traffic accidents and losses.For auto industry,ADAS could help to build the basis of Self-Driving in the future.As one of the main functions of the system,the research of preceding vehicle detection and collision warning is very important.