Research On Vehicle Distance Measurement Based On Lane Line Detection

Popularization of automobile brings people a lot of convenience, but also caused a lot of serious traffic problems. Intelligent vehicle is an excellent option to effectively solve those problems because of its highly intelligent advantages. The vehicle detection and distance calculation technology is fundamental to the realization of intelligent security system for vehicle driving safety. It allows intelligent vehicle for automatically detection of front vehicle and calculation of the distance, reaching the collision warning purpose.Basically, the collision warning system comprises of four processes:image collection and preprocessing, lane line detection, target detection, and distance measurement. The paper provides a more simple but effective preprocess algorithm: cut the upper40%area of the image, image gray, median filtering, edge detection. The Area of Interests can be obtained by detecting and fitting the left and right lane line. The outlines of the new algorithm are following:using the width of normal lane line as a threshold to remove most of interference points, making identifying edge points and fitting the lane lines precisely, due to the significant gray scale difference between the car and the road, the profile of the vehicle can be recognized. Then the two-dimensional coordinates of the vehicle in front in the image can be obtained. Eventually, using the formula for measuring distance that has been developed by ways of the transformation from2dimensions to3dimensions, along with the necessary parameters provided by camera calibration, the distance can be calculated.The results of the experiment has illustrated that the system successfully achieve desired functions:accuracy and real-time requirements. The last chapter of this paper issues several questions that need further optimization and research to this system.