Abstract:Lane detection based on computer vision is a key technology of Automatic Drive System for intelligent vehicles. In this paper, we propose a real-time and efficient lane detection algorithm that can detect lanes appearing in urban streets and highway roads under complex background. In order to enhance lane boundary information and to be suitable for various light conditions, we adopt canny algorithm for edge detection to get good feature points. We use the generalized curve lane parameter model, which can describe both straight and curved lanes. We propose an improved random sample consensus (RANSAC) algorithm combined with the least squares technique to estimate lane model parameters based on feature extraction. Experiments are conducted on both real road lane videos captured by Tongji University and Caltech Lane Datasets. The experimental results show that our algorithm is can meet the real time requirement and fit lane boundaries well in various challenging road conditions.

Lane Detection Method Based on Improved RANSAC Algorithm