All-in-One YOLO Architecture for safety Hazard Detection of Environment along High-Speed Railway
Ping Chen, Yunpeng Wu, Yong Qin, Huaizhi Yang · 2022 Global Reliability and Prognostics and Health Management (PHM-Yantai) · 2022
The safety hazards of environment along high-speed railway are paramount importance for railway transportation safety. The (UAV image)-based intelligent inspection method for environmental hazard is an effective alternative to manual inspection. This paper proposed an All-in-One YOLO architecture (AOYnet), which assembles semantic segmentation and object detection in two parallel tasks, aiming at eliminating the low- efficiency issue caused by single-task detector. Finally, the experiments evaluated on the Railway Surrounding Environment Dataset acquire 0.733 of mAP on detected objects and 0.972 of IoU on segmented objects, which outperform state-of-the-art methods. Thus, the proposed method is efficient and feasible for the detection of environmental hazards along high-speed railway using UAV Images.