On-drone analytics · Urban mobility
Aerial Traffic Patrol
Aerial Traffic Patrol explores how drones can observe road traffic and turn aerial video into useful vehicle detections, trajectories, and speed estimates.
01 / Research frame
The problem
A moving aerial camera sees vehicles at changing scale and angle while the platform itself is in motion. Reliable speed measurement also requires a grounded reference rather than image-space motion alone.
The approach
The work combines computer-vision methods customized for aerial traffic footage with field data that pairs high-resolution drone video and radar-gun ground truth. δ-ATP builds on the earlier D-AirPatrol architecture.
Contributors
- Jiaxin Du
- Shilong Lei
- Chunyi Peng
03 / Associated work
Associated publications
δ-ATP: Towards Fast and Reliable Aerial Traffic Patrol
Jiaxin Du, Chunyi Peng · ACM Drone Systems and Applications Workshop (DroNet), co-located with MobiSys
D-AirPatrol: A Dual-Layer Architecture for Traffic Patrol From the Sky
Jiaxin Du, Shilong Lei, Chunyi Peng · 30th Annual International Conference on Mobile Computing and Networking (ACM MobiCom)