Vision-based Vehicle Relative Distance And Speed Detection Method

With the high-speed and intelligent life,assisted driving has evolved into a non-negligible part of traffic driving.If the vehicle can automatically detect whether it deviates from the lane line during driving,this can greatly reduce the probability of accidents,especially for long-distance driving Drivers who are fatigued are even more important.This paper proposes a method for judging the safety of the vehicle while driving by the relative speed of the two vehicles and the relative distance change during the driving process of the vehicle.The main work and innovations included in this paper:1.In this paper,the lane line detection process is mainly divided into two parts,which mainly include image preprocessing and lane line detection.The traditional image pre-processing is to capture the image of its area of interest first to reduce its processing space.Based on this,this paper improves it to obtain its dynamic region of interest(DROI),and uses the grid method to divide its initial area of interest,which retains and removes pixels,dynamically acquires a small number of windows in real time,and uses the output of each frame as the input of the next frame,which can further reduce its processing space.A lane line detection algorithm combining Hough transform and Catmull-Rom spline curve is proposed,which overcomes the shortcomings that traditional lane line detection methods cannot effectively detect curved lane lines.2.Analyze the deficiency of traditional vehicle recognition,that is,the detection effect is not good under the condition of weak light and obstructed objects.An improved vehicle recognition method based on Haar features and Ada Boost classifiers,extended Haar features,and performs feature compression on Haar features are proposed.Enhanced the weak classifiers in Ada Boost classifiers and further enhance Strong classifier in Ada Boost classification.3.Based on the relationship between the established pixel coordinate system and the world coordinate system,the method of measuring distance in the ideal state and the method of distance measurement in the case of camera shake and pitch angle during vehicle driving are analyzed.The frame difference method and the method of superimposing multiple frames of images are used to obtain the relative speed of the vehicle,and several situations that may cause experimental errors are analyzed.4.The traditional calculation method of safe distance is introduced.An improved vehicle collision avoidance safety distance model is established,and the safety of vehicles in different environments is analyzed according to the relative distance and relative speed of the two vehicles.