| At present,the questions restricting the development of road traffic system mainly focus on traffic accidents,traffic pollution and traffic congestion.In the car began to widespread use of electronic auxiliary equipment,mainly realize the auxiliary function of electronic stability control,lane departure warning,as well as the emergency brake and so on.But to reduce artificial factors lead to the problem of traffic accidents,the vehicle used by electronic auxiliary system cannot fully achieve the purpose of intelligent auxiliary driving,also cannot use accident warning,as well as risk assessment form,to reduce the traffic accident.Therefore,in order to realize the application of auxiliary driving system or unmanned technology to reduce traffic accidents,the purpose of one problem need to solve is to realize the real-time and reliable traffic environment perception.Aiming at these problems,this paper proposes a scan raster environment based on stereo vision and perception system,used to achieve accurate real-time road environment 3 d reconstruction.Paper main research content is as follows:(1)System mathematical model establishment and system calibration.According to the requirements of the visual field of perception of the front environment during road driving,the parallel binocular stereo vision model was selected as the visual model of the environment perception system.Since the camera coordinate systems can be converted to each other,a mathematical model is established by using the conversion relationship to realize the conversion of the world coordinate system to the image coordinate system.(2)The center point of the grating stripe is positioned.Aiming at the problems that the current common grating stripe center point extraction algorithm cannot adapt to the change of the grating stripe direction,combined with the actual situation of the stereo vision and scanning grating system and the characteristics of the grating stripe image,a kind of extraction of the grating stripe center based on the skeleton normal is constructed.Point algorithm makes it possible to accurately locate the center point of the grating stripe even when the direction is very complicated.(3)Image stereo matching and reconstruction.By comparing the matching performance of different algorithms,combined with the actual application scenarios of the environment perception system,a fast matching algorithm based on epipolar line constraints is proposed to achieve fast and accurate matching of the left and right raster images of the environment perception system.Analyzes the influencing factors of road environment perception,design the overall research of environmental perception system,set up a software and hardware system integration test platform together.From the end of the experimental results,this paper put forward based on stereo vision and the path of raster environment perception system can realize the accurate 3 d reconstruction of the goal,improves the environment perception system security and reliability,the unmanned for unmanned decision-making system provides a complete,reliable,accurate data to support. |
