Research On Augmented Reality Navigation Technology Based On Vehicle-mounted AR-HUD

With the development of information technology,people have become accustomed to using mobile phones for navigating when driving.However,the driver may divert their attention from controlling the vehichel with the line of sight being transferred from road to phone,which will bring safety risks.The Vehicle-mounted Augmented Reality Headup Display(AR-HUD)system ensures the driver driving the vehicle normally while provides driver more intuitively relevant navigation information by projecting the information integrated the information of normal phone navigation with the lane infomation obtained from environmental awareness to the driver's front vision.At present,the augmented reality navigation technology based on vehicle-mounted AR-HUD has become a research hotspot in the field of automobile intelligent technology.However,there are still many key technical problems to be solved in order to realize industrialized application.This thesis study the registration of virtual and real information,detection of lane line,estimation of road relative slope and AR-HUD augmented reality navigation system.The main research contents as follows:1.This thesis aims at solving the problem of the real-time and accurate registration and fusion of electronic map navigation information and the real world information such as driving path,lane line.The virtual image model of vehicle-mounted AR-HUD system is established.And the engineering realization method of relevant parameter calibration,measurement and registration of virtual and real of the AR-HUD system is studied.On this basis,the matching,alignment and prompt of navigation virtual markers and realistic targets of the vehicle road are realized.The effectiveness of the proposed method is verified by vehicle experiments.2.To solve the problem that AR navigation system needs to enhance the accuracy and sense of reality of driving guidance by detecting the current road information such as lane line,lane curvature and relative slope in real time,this thesis proposes a lane line detection algorithm based on monocular vision using dynamic ROI and second location.By using the quadratic curve model and the least squares fitting,the dynamic estimation and prediction methods of lane trend,curvature and other lane elements are proposed.And the concept of the projection curve of the lane curve based on monocular vision detection on the horizontal coordinate plane of the vehicle coordinate system through camera perspective transformation is proposed for the first time.And a real-time estimation algorithm of the relative slope of the lane is proposed.The correctness and effectiveness of the algorithm are verified by simulation and real vehicle experiments.3.For the actual engineering application,a AR navigation application system based on AR-HUD is designed.The framework of the system software design,software module design and UI design are finished.And a prototype system with the route guidance,driveway tip and AR navigation functions based on the ARM architecture processor and Linux vehicle platform is developed.The real car test shows that the algorithm in this paper plays an important role in improving the functions related to AR navigation,enhancing the fusion effect of virtual and real road scenes,and improving the AR-HUD user experience.