The tello-python from niusha-tw

Python 3 version of https://github.com/dji-sdk/Tello-Python

sudo apt install python3-venv
python3 -m venv testTello
source testTello/bin/activate

Connect your machine to the Tello via WIFI
Run the code and fly Tello autonomously (it would take off, fly around and land by its own)

python3 fly_tello.py

Install necessary libraries if you need image from Tello (opencv) and show it in your web browser (flask).

pip install opencv-python Flask # That is it. No other dependent libraries

Run the code and visit http://127.0.0.1:9999/stream.mjpg by using your browser (keep refresh the page if you have problem)

python3 see_from_tello.py

Tello --[h264-encoded chunks of frame]--> UDP server on my PC --[h264-encoded frame]--> TCP server on my PC --[h264-encoded stream]--> OpenCV --[image]--> Flask
We cannot get a video stream from Tello so we cannot use OpenCV directly.
- When I worked with raspberry pi, its camera utility can set up a tcp server then OpenCV can capture the h264 stream from it directly. But Tello is different because tello is not going to setup a streaming server (and it is not programmable to do that).
I was searching for a full-proof solution to decode locally buffered h264/H264/H.264 stream and get the image frame by frame.
- Tello is a client who sends out chunks of h264-encoded video stream to a udp server set up on my PC. It is up to me to collect those chunks and deal with it.
- I am familar with OpenCV but I don't know how to work with a locally buffered stream.
- In Tello_Video.ipynb, I demonstrated a solution by setting up a tcp server to rebroadcast the buffered stream. Yes, however we get the raw stream, we can feed it to a tcp server. OpenCV will capture the tcp stream and do the real-time decoding frame by frame.