An evolutionary computation approach for lane detection and tracking

An evolutionary computation approach for lane detection and lane to tracking is proposed. The particle swarm optimization (PSO) is used to detect and track the lane position of the road image which is obtained in real time. First, this method converts the RGB color image to the YIQ color space to isolate the gray-scale information in a color image. Second, a spatial filter is utilized to extract the available lane features from the region of interest (ROI), and then PSO is used to search for the correct lane position in the gray-scale image. In order to enhance the effectiveness of PSO in lane detection, some representative road images are selected from the image database as the training patterns for tuning of the structure and parameters of PSO-based detection. For testing the performance of the proposed method, a road image database with more than 8000 actual road images has been tested for evaluating system performance. The experimental results show that the performance and efficiency of the proposed method can achieve the objective of real-time lane detection and tracking for the real-road conditions.