raspbian-EdgeTPU

This Docker image includes the libraries and packages to work with the Coral USB Accelerator (Google). The based image is balenalib/raspberrypi3-debian:buster.

The image contains the following packages/libraries:

• Python 3.7.3
• NodeJS 12.13.1
• Python Packages:
  • numpy, matplotlib, pil, zmq
  • supervisor, tornado, picamera, python-periphery
  • jupyter, cython, jupyterlab, ipywebrtc, opencv-python
  • google-auth, oauthlib, imutils
• Other libraries included (check the Dockerfile)

A Jupyterlab is available (https://<ip-address>:8888) in which you can write code to process images obtained e.g. from the Pi camera. Moreover, two examples using the Coral USB Accelerator are included:

• webcam_obj_detector_picamera.py: detects and classifies objects on the fly processing the images taken with the Pi camera. The streaming images are available over http (http://<ip address>:8080). It uses the picamera library to get images from the Raspberry Pi Camera (only CSI connector).
• webcam_obj_detector_opencv.py: detects and classifies objects on the fly processing the images taken with the Pi camera. The streaming images are available over http (http://<ip address>:8080). It uses the opencv library to get images from the camera (USB port or CSI connector).
• teachable_jupyter.ipynb: a notebook to re-train the tflite models to include your objects for image classification.

To run the container type the following on a Raspberry Pi terminal:

docker run -d --privileged -p 25:22 -p 8080:8080 -p 8888:8888 -e PASSWORD=<<JUPYTER_PASSWORD>> --restart unless-stopped -v /dev/bus/usb:/dev/bus/usb  lemariva/raspbian-edgetpu

You need to activate the camera interface using sudo raspi-config to use live images of the Raspberry Pi camera.

More Information

More information about the repository can be found on the following links:

• #Edge-TPU: Coral USB Accelerator + rPI + Docker
• #Edge-TPU: Hands-On with Google's Coral USB accelerator

Licence

• Apache 2.0