Research On Highway Sight Distance Safety Evaluation Method Based On Reliability And Monocular Vision System

With the increasing number of motor vehicles and the production of new energy vehicles,road traffic safety issues have become increasingly prominent,including congestion and traffic accidents.Sight distance is a key indicator to ensure driving safety,and it is also one of the design indicators in highway route design specifications.Sight distance is related to the safety and speed of driving.Ensuring that drivers can obtain sufficient sight distance at various points along the highway is the key to ensuring road safety.Therefore,the calculation of sight distance of road driving and the inspection of whether the visual distance meets the requirements of driving safety is a key content of highway design.It is of great significance to detect the highway line of sight in a timely and accurate manner,find the section with insufficient line of sight and take improvement measures for improving the safety of highway driving.This paper studies the road lane line detection technology and monocular vision system,establishes a line-of-sight model based on monocular visual ranging,and tests the line-of-sight of each point along the highway.Combined with the reliability and safety evaluation method,The safety evaluation of the driving sight distance of each point along the highway is carried out and improvement measures are proposed.First of all,to realize the detection result of highway line-of-sight based on the monocular vision system,the first stage is image processing.The object of image processing is the image captured by the camera with the original road information.Image processing includes road image preprocessing,image edge detection and lane line detection in three steps.This paper uses color image grayscale,equalization grayscale histogram and edge detection with Canny operator to preprocess road images.Image segmentation adopts the iterative threshold segmentation method to achieve image enhancement for the image whose gray histogram is equalized.Secondly,three basic assumptions are put forward: the width of the structured road is basically unchanged,the road is flat,and the road model is a straight line model.The modelbased lane line detection method and Hough transform are used to detect the lane lines.Next,this thesis studies the world coordinate system,camera coordinate system,pixel coordinate system and image plane coordinate system,analyze the mapping relationship between the image plane coordinate system and the world coordinate system,established a geometric relationship,and calculates the lane line and pixel coordinate system in the world coordinate system.The mapping relationship between the inner lane lines,using the Zhang Zhengyou calibration method to calibrate the internal and external parameters of the camera,correct the camera lens distortion,etc.,to reduce the error.Using the single frame static image ranging geometric model and cubic B-spline interpolation,study the geometric calculation relationship between the pixel coordinate system and the world coordinate system,establish a monocular vision ranging model,obtain the quantitative relationship of the world coordinate system from the pixel coordinate system,and solve the problem according to the calculation model.Sight distance detection results.Finally,combined with the example of a national highway project in Guizhou Province and the reliability safety evaluation method,the line-of-sight measurement model is applied to the actual highway engineering,combined with the project video data,the line-of-sight test is carried out for each point along the highway,and the monocular visual distance measurement in this paper is used.The model is applied to engineering practice,and the safety evaluation of driving sight distance is carried out for the projects in the delivery and operation stage of highway engineering,and improvement measures are proposed.