Research Of Lane Recognition And Tracking In Haze Weather Based On Machine Vision

With the rapid increase of the number of cars,driving safety has been paid more and more attention,and the lane recognition and tracking algorithm has become a hot research field.The traditional lane recognition and tracking algorithm is mainly aimed at a good environment,but little research has been done on adverse environments such as haze weather.For the haze environment,the contrast of the roadway image roadway and lane is much lower than the normal environment,which brings difficulties to the identification and tracking of the lane.This paper studies the lane recognition and tracking algorithm in haze environment.The content include the following aspects:(1)Research on imaging principle and image defogging algorithm in haze environment.Aiming at the problem of poor real-time performance of traditional dark channel dehazing algorithm,a dark channel dehazing algorithm based on morphological erosion operation is proposed.Compared with the traditional dehazing algorithm,the algorithm can achieve fast and accurate image defogging.(2)For different types of lane,different recognition algorithms are used in this paper.For straight lanes,the straight line model is chosen.For the problem of inaccuracy and poor real-time performance of traditional Hough transform in lane recognition.A lane recognition algorithm based on polar angle constraint and classification discriminant Hough transform is proposed.The experimental results show that the algorithm not only recognizes the lane on both sides of the lane,but also improves the real-time performance.For the curve,the parabolic model is chosen.Firstly,the particle swarm algorithm has premature convergence.The chaotic theory is used to improve the particle swarm optimization algorithm.Then the lane objective function is designed by using the difference between the lane and the road gray value.Finally,the chaotic particle swarm optimization algorithm is used to optimize the parameters in the objective function.The experimental results show that the proposed algorithm can not only identify the curved lane in different situations,but also meet the requirements of real-time.(3)In order to improve the real-time performance of lane tracking,this paper uses the recognition result of the previous frame to predict the next frame image based on the difference of the lane lines between adjacent frames.The lane is tracked by the chaotic particle swarm algorithm.The experimental results show that the method can track the lane quickly and accurately.(4)The lane recognition and tracking algorithm in the haze environment designed in this paper has important theoretical significance and application value.