| In recent years,with the rapid economic development,people’s living standards have been continuously improved,and people’s travel methods have also changed.Self-driving travel has gradually become the first choice for travel,but the resulting traffic congestion problems are becoming more and more serious.Accidents occur frequently.The existing transportation system is difficult to fundamentally solve these problems.In order to improve and solve the problem of traffic congestion,intelligent transportation systems and autonomous driving technology have become hot topics for scholars at home and abroad.Among them,vehicle self-positioning is the basis and key technology in intelligent transportation systems and autonomous driving.It is extremely important for driving safety and driving decisions.In view of the limitations of the existing vehicle self-positioning algorithm due to the driving environment,some algorithms have high complexity and high cost,this thesis proposes a new position estimation algorithm based on the new standard road signs for the vehicle monocular camera,which realizes the complexity of urban traffic accurate and efficient vehicle position estimation in the road environment.First of all,in view of the defects of the existing road indicating feature points that are not obvious,a novel standard road sign with obvious characteristics and simple structure is proposed and designed.Secondly,a database consisting of the size of standard street signs,information of control points,and the number of lanes on the road where the standard street signs are located,the width of the lanes and the direction of travel is builted.Next,a vehicle-mounted monocular camera is used to collect a road image containing a new type of standard road signs,and the control area of the standard road signs is extracted through image pre-processing segmentation,and the centroid point of the control area is obtained as the control point for subsequent calculation.Then use this control point and the control point information pre-stored in the database to calculate the plane projection transformation matrix,and use this matrix to register the reference street signs in the database to the standard street signs in each frame of the image to verify the extracted control Calculation accuracy of points and plane projection transformation matrix.Finally,after solving the morbidity of the equation system,the matrix is decomposed to obtain the position information of the vehicle in the coordinate system with the center of the standard street sign as the world coordinate origin,and compared with the actual measured data.The experimental results show that the detection rate and accuracy of the new standard road signs proposed in this paper are significantly higher than the existing road signs.The coordinate error calculated by the centroid point is within one pixel error.The calculation error of the vehicle position estimation result is within 0.5 meters from the fixed position of the standard street sign 100,the positioning accuracy within 30 meters can reach 0.1 meters,and the positioning accuracy within 20 meters can reach 0.05 meters.The method proposed in this paper is low cost,simple and reliable.Using the on-board monocular camera and standard street signs,the vehicle’s position information and lane information are accurately calculated,and the vehicle’s precise self-positioning in complex traffic sections is achieved.In this thesis,first elaborates the research background of Intelligent Transport System(ITS)and vehicle self-localization technology and the current status and progress of domestic and foreign research,introduces the research content and algorithm flow of this article,and then introduces the preprocessing process of the input image of the car monocular camera,and derives in detail The calculation process of the vehicle position parameters obtained from the plane projection transformation matrix is solved.Finally,the experimental analysis and resultsdiscussion are carried out. |
PDF Full Text Request