A Vision-based Lane Departure Warning Framework

Lane departure warning is an essential function of driver assistance systems. In this paper, we propose a vision-based lane departure warning framework, which can determine whether the vehicle deviates from the lane only with images as input. Our framework combines the deep learning method and traditional line detection algorithm. Specifically, the framework consists of a Lane Boundaries Localization module, a Lane Boundaries Generator module, and a Departure Warning module. The Lane Boundaries Localization module is a deep neural network for finding the area containing lane Boundaries. Then the Lane Boundaries Generator module performs backbone detection and line detection to generate lane boundaries on the located area. After identifying the lane boundaries, the Departure Warning module can quickly determine whether the vehicle deviates from the lane. Additionally, we implemented and deployed our framework to an Android platform and tested its running speed.