Dynamic Safe Area Detection Method for Autonomous Vehicles in Hub Airport

The application of autonomous driving technology will improve the level of airport autonomy and intelligent operation. In order to improve safety, autonomous vehicles need to accurately detect what portion of area inside the lane is free from obstacles. This paper proposes a safe area detection method, which uses a safe threshold to combine the lane detection model and the SSD object detection model to mark the safe area for airport autonomous vehicles. Finally, in the built airport simulation platform, we mainly compare the proposed method with the FCN area detection method. The simulation results show the proposed method improves accuracy rate by 7.91%, and speed by 17.90%, which can meet the requirements of real-time detection in hub airport.