Lane Departure Detection and Early Warning System for Intelligent Vehicles Based on Yolov3 Algorithm

IV (intelligent vehicles) lane departure detection and early warning system can improve driving safety, reduce accidents caused by distracted or tired driving, and it is also an important part of automatic driving technology, which helps to achieve greater safety. In this paper, the detection and early warning of IV lane departure is taken as the research object, and a detection and early warning system of IV lane departure based on YOLOv3 algorithm is constructed, so as to realize end-to-end detection with actual road images as information. The network structure of YOLOv3 algorithm is improved, and the network is simplified to 101 layers, which effectively reduces the risk of missed detection of lane small targets in complex background. Combining Hough transform and YOLOv3 algorithm, the lane detection in the image is realized, and good results are obtained. The research results show that the average detection time of YOLOv3 network with improved network structure is shortened by 2.04 ms, the mAP is increased by 5.04%, and the average miss detection probability is reduced by 6.37%, which effectively avoids lane miss detection, and the detection accuracy is significantly higher than other algorithms. The algorithm has strong robustness in complex road environment and can meet the requirements of accuracy and speed of IV lane departure detection and early warning in the actual intelligent driving process.