CALL FOR PAPERS

Drones can provide a wider field of view, high mobility, and flexibility for monitoring and analyzing traffic flows and safety conditions than security CCTV cameras. In the case of a perpendicular viewing angle to the ground, minimal occlusion can occur and make vehicle tracking or people counting be easier. However, due to its observation far from the ground, limited battery time, and bandwidth, this solution should be edge-based and have a reasonable recognition rate in small object detection / semantic segmentation. Furthermore, Much cheaper, lightweight quadrotors are available on the market, convenient to control through Wi-Fi by smartphone or laptop. However, the risk of a small quadrotor crashing obstacles is pretty high, especially in cluttered unknown environments. Therefore, perceiving and avoiding obstacles is necessary for a small UAV and is also highly challenging tasks, especially for a quadrotor only depending on a monocular/stereo camera to sense and navigate in cluttered unknown environments. To deal with the above problem, the integration of computer vision, sensing technologies, control systems, signal processing, communication, hardware design, and machine learning into a drone is inevitable for correctly understanding the environments and making correct and safe decisions. The special session will cover all these topics. The topics of interest include, but are not limited to :

  • Small / Density / Real Time Object detection and semantic segmentation
  • Visual object tracking in unmanned aerial vehicle (UAV) with camera motion
  • Monocular / Stereo Depth Estimation for UAV Obstacle Avoidance
  • Multi-modal learning for UAV intelligent System
  • Reinforcement learning for UAV self-control
  • Hardware design for UAV system
  • Scene Recognition for Complicated UAV Images
  • UAV image dataset construction
  • Crowd Counting on the UAV system with high density objects

Maximum Length of a Paper

Each full paper should be limited to 6-8 pages (6 pages limit + references).