A Lane Departure Warning System Based on Machine Vision

Lane departure warning system based on machine vision is a human decision-make like solution to avoid lane departure fatalities with low cost and high reliability. In this paper, the model of vision-based lane departure warning system and the realization is described at first. Then the method of lane detection is illustrated, which is composed of three steps: image preprocessing, binary processing and dynamical threshold choosing, and linear-parabolic model fitting. After that, the solution of how to perform the departure decision-making is proposed and demonstrated. Unlike the usual TLC (Woong Kwon et al., 1999) and CCP (Risack et al., 2000) methods, the angles between lanes and the horizontal axis in captured image coordinate are used as the criterion for lane departure decision-making. At last the experiments are implemented by use of all the steps; the results indicate the efficiency and feasibility of the solution.