Tello Edu swarm with multi-stream support and object detection

This was a student group project aimed to evaluate the capabilities of Tello Edu drone swarms. We achieved:

• Manually controlling the drones in a swarm
• Displaying streams from all the drones in a swarm with real-time object detection
• A mapping system which allows:
  • the drones to avoid collision with each other.
  • the detected object to be drawn on the map

Multiple streams

For multiple streams you will need separate network card/usb wifi dongle with static ip for each drone.

Object detection

easy-yolov7 is used for object detection.

Install all easy-yolov7 dependencies:

pip install -r requirements.txt

Install cython-bbox for Windows:

pip install -e git+https://github.com/samson-wang/cython_bbox.git#egg=cython-bbox

For better performance set "DEVICE" in swarm.py to "gpu" (requires CUDA-Enabled GPU).

Mapping

The drones' coordinates on the map are calculated from user inputs.