Intelligent real-time lane detection for vehicles in urban streets

An efficient method of lane detection is proposed to improve robustness and efficiency of lane recognition in urban streets. The top view of interest area is established using inverse perspective mapping. Lane detection based on a curve detector has good robustness and can identify lane accurately. It also can extract the yellow lanes in real time with the HSV space. Experiment results show that the algorithm has good performance under different weather conditions and lane models. This method can be applied to the intelligent vehicle.