Design Of Intelligent Signage System Based On Road Information Detection

Road signs play an important role in traffic safety.It can guide the direction and warn the danger,provide the driver with accurate road indication information,avoid causing major traffic accidents.Aiming at the problem that traditional road signs can only use fixed words and symbols to convey information to traffic participants unilaterally and cannot detect road conditions in real time.This thesis proposes to design a sign system based on road information detection,which can detect road water and traffic flow information in real time and issue warnings to remind drivers to move forward carefully.According to the perception,display and monitoring requirements of signs on road information.The system is divided into four parts,including water detection unit,traffic flow detection unit,sign display unit and remote monitoring platform.(1)The water detection unit is used to detect road water conditions.Through the design of ultrasonic driver in sound speed correction and road disturbance filtering,the stability of the detection process is enhanced,and vehicle flameout and other traffic accidents caused by the deep water accumulation on the actual road are avoided.(2)For the traffic flow detection unit,this thesis proposes a traffic flow detection model based on CAMShift multi-feature fusion and Kalman prediction in combination with the visual background extractor algorithm.Vibe algorithm was used to extract target vehicles in the video,and the target detection algorithm is further improved in the aspects of "ghost" elimination,contour filling,light filtering,and shadow processing.Aiming at the problem that the CAMShift algorithm fails to track when it is blocked by color and occlusion in practical application scenarios.This thesis establishes a multi-feature vehicle tracking model.Multi-feature fusion includes tracking the vehicle’s paint,edge,and texture.When the tracking target is not blocked,the CAMShift algorithm calculates the target position and updates the Kalman filter parameters.When there is occlusion,Kalman is used to predict the current target position.(3)The label display unit is used to display road water information and traffic flow information to remind drivers to drive carefully.(4)Remote monitoring unit.Using NB-Io T technology to connect the signage system with the cloud platform,for uploading detected detected road water information and vehicle flow information.Finally,the whole system was tested,including hardware platform test,water detection unit test,traffic flow detection unit test,Internet of Things platform communication test.In the process of water detection unit test,experimental simulation test and actual environment test were carried out,and the detection error was verified to be within 3 cm.In the process of traffic flow test,the effectiveness analysis of the improved Vibe algorithm,the target tracking robustness test under different interference environments,and the vehicle tracking error analysis were carried out.Through comparison with different algorithms,it was proved that the proposed algorithm had strong robustness under different environmental interference.The accuracy of traffic flow detection is 92～97%.Data can be transferred between the system and the cloud platform,and the overall system operation meets the expected goals.