Research On Drive Maintenance System Based On Binocular Vision

With the continuous improvement of people’s living standards,cars have become the main means of daily travel,and the popularity of cars brings not only the convenience of transportation,As an important part of vehicle active safety system,driveway maintenance system has received widespread attention in the field of scientific research.Therefore,the study of lane holding system is of great practical significance.The main research content of this paper is summarized as follows:1.In the preprocessing of the image,firstly,the camera is calibrated according to the imaging principle of binocular vision,then the acquired image is preprocessed :firstly,the division of the static region of interest is carried out,and then the preprocessing image is gray-scale by using a weighted average algorithm.2.Lane line processing: there are two common methods for extracting lane lines through feature points at present: two kinds of the Hough transform and the least square method.The Hough transform is more suitable for the road with a straight line model,so if the lane sign line has a degree of curvature,A quadratic or more complex curve model is needed for the curved part.And in terms of the overall slope of the driveway,The linear results extracted by the least square method can describe the direction of lane extension more accurately.The parameters of the model and the complexity of the algorithm should be considered in order to select the best scheme.3.The establishment of lane deviation early warning system: Firstly,the models are established: vehicle model,driver model and vehicle road model.Then the algorithm of the controller is designed.Because the input of the system is the steering wheel angle,the lateral acceleration of the vehicle is controlled according to the knowledge of the steering system.This paper adopts the fuzzy PID control method.Finally,the simulation experiment platform of the algorithm is built,and the feasibility of the proposed algorithm is proved by the analysis of the experimental data.