Lane Detection in Autonomous Driving: A Comprehensive Survey of Methods and Performance

Lane recognition is an essential task for autonomous cars and driver-aid systems. However, the currently used algorithms for lane detection may not perform well in certain situations, such as rainy weather, fog, and especially at night when lanes are not visible. A recent study presents a comparative analysis of five-lane detection algorithms, including color thresholding, gradient-based approach, edge detection, Hough transform, and the convolutional neural network (CNN). The study provides a brief overview of each algorithm in a table and assesses their effectiveness. The analysis concludes that the CNN algorithm outperforms the others in terms of both accuracy and flexibility.