jetson-tello

Utility code for using the NVIDIA Jetson and tello-asyncio to interact with the Tello EDU drone.

The primary function so far is to pipe video frame data from the drone through to neural networks running on the Jetson, typically for object or face detection.

Most importantly, the frame data analysis can take as much time as it needs. Frames are skipped if needed to keep up with the live video stream.

Created for my autonomous drone hobby project, drone-braain, which is very much in its infancy.

Package jetson-tello on PyPi.

Prerequisites

There are two prerequisites that require manual installation:

• NVIDIA’s jetson-inference project, following these instructions to build from source and install.

• My fork of h264decoder. This is identical to the original repo apart from building with the slightly old version of CMake (3.10) available on the Jetson.

• Python package: jetson_tello
  • jetson_tello.app
  • jetson_tello.coco
  • jetson_tello.video
• Code examples
  • Face detection
  • Face and object detection
  • H.264 frames to NumPy arrays
  • H.264 frames to CUDA
  • Object detection