The global civil aviation is booming as air passenger volumes continue to grow at a rapid pace, as well as the number of flights and routes. The upward trajectory in air traffic adds to the most pressure on airports. Airports need to get smarter in order to meet the operational and capacity challenges, especially for capacity constrained airports. Airport ground video surveillance is crucial for intelligent airport applications, such as visual conflict alert and docking guidance. Change detection (or moving object segmentation) and object tracking are the foundation of video surveillance, but usually fails to get satisfactory results in airports. It is necessary to establish a change detection dataset and an object tracking dataset, and then develop specific algorithms for airport ground video surveillance.
An Airport Ground Video Surveillance benchmark for change detection (AGVS-S) and an Airport Ground Video Surveillance benchmark for Object Tracking (AGVS-T) are presented on this website. Considering that the airport ground is semi-militarized zone, we gain access to the ground of a huge airport through cooperation with the Civil Aviation Administration of China (CAAC). The two datasets are available here:
1. Airport Ground Video Surveillance for Change Detection (AGVS-S);
2. Airport Ground Video Surveillance for Object Tracking (AGVS-T).
3. Airport Ground Video Surveillance for Video Semantic Segmentation(AGVS-VSS).
4. A Moving Target Recognition Dataset for Airport Ground Video Surveillance(AGVS-R).