关键词: YOLO, 障碍物检测, 深度学习

Abstract: In this paper, an obstacle detection method based on YOLOV3 algorithm is proposed for obstacle detection of medium-low speed trains in transit. First, the deep learning algorithm is used to effectively identify obstacles in the scene. Then, the edge detection algorithm based on Freeman chain code is used to extract the edge of the train track, so as to determine whether the obstacles affect the traffic and give a warning to the driver. At the same time, the YOLOV3 network data set is expanded by means of transfer learning, so as to achieve the purpose of improving the target obstacle recognition accuracy by use of this method in a specific scene. The experimental results show that this method has high applicability and is a convenient and efficient obstacle detection method for trains in transit.

Key words: YOLO, obstacle detection, deep learning