UAV2UAV-dataset (referred to as "U2U") is an airborne visual single object tracking dataset with a total of 54 videos and about 25,000 frames The targets of the U2U dataset are fixed-wing UAVs, which including the common level flight, large slope turn under the overhead view and stunt flight actions, such as vertical climb and lateral roll. The ground background includes relatively monotonous wasteland and grassland, as well as urban buildings, sports fields and other complex scenes.
The dataset can be downloaded https://pan.baidu.com/s/1QWndfxUkBa4YU0FEf0Z8jw (pwd: 9oe9) or https://drive.google.com/drive/folders/1ZPmR2jIAetkypgrc6VurUXxqE7_LUxFK?usp=sharing.
The first frame annotation is stored in [left top weight height].
Yong Wang, Zirong Huang, Robert Laganière, Huanlong Zhang, Lu Ding, A UAV to UAV tracking benchmark, Knowledge-Based Systems.
We would like to thank the following contributors for working on the collecting, annotations and conducting the quality checks for this dataset.
SYSU Unmanned Aerial Vehicle Association（中山大学无人飞行器协会）
Yuguang Chen
Zirong Huang
Jingming Yan