Lane Detection based on Deep Learning

Jian Chen ； Dan Yu

《International Core Journal of Engineering》 7卷6期 (2021/06) Pp. 639-652

https://doi.org/10.6919/ICJE.202106_7(6).0075

摘要

The detection identification of the lane line is an important part of automatic driving and advanced driving assistance. It is well known that the computer is not as simple as human beings for the identification of lane lines. Although there are many lane detection processing methods based on traditional computer vision. However, these processing methods are not ideal in the state of road conditions or poor illumination conditions. This paper discusses a method of detecting recognition based on deep learning. A convolutional neural network (CNN) is used, a neural network that is good at processing image data. The idea of the project is to make the network with the ability to calculate the road polynomial coefficient, and use these coefficients to predict the lane cable. Finally, draw the current driving lane based on the predicted road line. This is similar to a regression problem. And use average variance to narrow the loss of training. After the final experiment, the model of this paper has a good performance in complex circumstances, combining time-effective, high accuracy, and high robustness.

關鍵字

Lane Identification ； Convolutional Neural Network ； Deep Learning

參考文獻

Kunihiko Fukushima. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position [J]. Biol. Cybernetics 36, 193 202 (1980).

Yim Y U, Oh S Y. 4“hree-feature based automatic lane detection algorithm (TFALDA) for autonomous driving[I]. IEEE "fransactions on Intelligent Transportation Systems, 2003, 4(4): 215-225.

Li, Mu, et a1. Efficient mini-batch training for stochastic optimization[C]. Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 2014.

Gurghian, T. Koduri, S. V. Bailur, K. J. Carey, V. N. Murali, DeepLanes: End-To-End Lane Position Estimation Using Deep Neural Networks. CVPR Workshops, pp. 38-45, 2016.

J. Li, X. Mei, D. V. Prokhorov, D. Tao, Deep Neural Network for Structural Prediction and [9] Lane Detection in Traffic Scene. IEEE Trans. Neural Netw. Learning Syst., vol. 28, no.3,690-703, 2017