Lane Detection Based On The Principle Of Human Visual Cognition

Nowadays,with the rapid development of global economy and the popularization of family cars to thousands of households,great changes have taken place in the mode of human transportation.The benefits are self-evident,but at the same time,the traffic environment and ecological environment are getting worse and worse.In order to solve these problems,we must adopt modern methods to enhance traffic safety,people begin to explore efficient lane detection system.However,the robustness and accuracy of the existing algorithm need to be improved,especially in the case of complex and diverse road environment,the accuracy of the result recognition still has some errors.In order to solve the above problems,considering that human vision is the most important way for drivers to obtain the surrounding road conditions,this thesis integrates the cognitive principle of human vision,and explains and designs the lane recognition algorithm according to the characteristics of human vision.The main contents and innovations of this thesis are as follows:(1)This thesis explores the way of the driver’s binocular vision system to the road environment,expounds the cognitive principle of human eye vision,and analyzes the influence of the principle of the driver operation during driving.(2)The process of perception and processing of vehicle information by human vision is simulated by computer,combined with the cognitive model of human vision,the traditional image preprocessing method is improved,includes: a road vanishing point detection algorithm based on probability voting method is proposed to estimate the vanishing point position,and according to the vanishing point,the parameters of perspective transformation are determined;wavelet transform and Retinex theory are used to enhance image details;the method of dynamic recognition of region of interest is proposed;it solves the problem that the traditional Canny edge detection needs to set the threshold manually.(3)The algorithm of lane recognition is analyzed and designed,it includes: the principle of image histogram is used to locate the starting point of the lane line,and the sliding window technology is used to track the lane line;the complete lane line is fitted by quadratic curve function,and then the validity of lane line curve fitting is verified.(4)Several performances of the proposed algorithm are tested.A number of data sets are used to test the algorithm,and its accuracy and time-consuming are analyzed;And algorithm is compared with other algorithms in cited literature.(5)Design and implement the lane recognition system,modularize the system function,and test the function and performance of the system.This exploratory thesis is based on the principle of human visual cognition,it explains and expounds the existing lane recognition technology from a new point of view,which is an innovation in the research of lane detection technology.At the same time,the traditional lane detection algorithm is improved from the principle,this thesis makes a breakthrough in thought,which is improve the accuracy of the existing lane detection algorithm.The algorithm designed in this thesis is the practical application of the principle of human vision cognition in the field of lane detection,which provides a new idea and perspective for the construction of advanced driving aid system.