| Currently, the accuracy and detection speed of lane detection in complex driving scenarios by the existing automatic driving systems are not high. In order to improve the accuracy and speed of lane keeping function in autonomous driving vehicles, this paper proposes a method of using convolutional neural networks for feature extraction, combined with a classification network, to achieve the classification of multiple lane lines. Test experiments under various conditions such as straight driving, turning, uphill and downhill driving, road bumps, and uneven lighting, show that the accuracy of the algorithm proposed in this paper can reach 95.14%, and the detection speed is also significantly improved compared to traditional mainstream algorithms. |