Abstract
With the rapid development of lane detection technology in the field of autonomous driving and intelligent transportation, its robustness and real-time requirements in complex environments are constantly improving. Although the traditional image processing method has a fast-processing speed, it has obvious limitations in complex scenes. Although the segmentation method based on deep learning has advantages in accuracy and adaptability, its high computational cost and large hardware requirements limit its practical application. To solve these problems, this paper adopts the structure-sensing depth lane detection algorithm based on global features. By introducing global features and larger receptive field, the algorithm simplifies the image segmentation process and improves the computational efficiency. We tested TuSimple and CULane datasets, and the results showed that the algorithm significantly improved the detection speed and environmental adaptability while maintaining high accuracy, especially in complex environments such as rainy, foggy and night. This paper further discusses the potential improvement direction of the algorithm in practical application, and lays a foundation for the future lane detection technology.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call