An Illegal Target Intrusion Detection System of Railway Based on Deep Learning and Hough Transform
Gangdi He, Nan Zhang, Xiaorun Li · 2021
Accurate detection of illegal target is important to ensure the railway safety. In this paper, an illegal target intrusion detection system of railway based on deep learning and Hough Transform is presented. This system employs Hough Transform to detect railway and YOLOv3 as well as SSD to identify target. Meantime, online hard example mining(OHEM) is adopted to solve the imbalance between simple samples and hard samples in dataset. In the detection process, results of railway and target are combined to distinguish whether there is an alarm or not. The experimental results show that the detection system has good accuracy and robustness.