Skip to content

Object detection live streaming using Coral USB Accelerator

License

Notifications You must be signed in to change notification settings

pengdev/coral_live_stream_demo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

85e713a · Sep 1, 2019

History

11 Commits
May 17, 2019
May 18, 2019
May 18, 2019
May 17, 2019
Sep 1, 2019
May 28, 2019
May 28, 2019
May 28, 2019
May 18, 2019
May 18, 2019
May 18, 2019
May 18, 2019

Repository files navigation

Coral object classification model live stream

Description

Stream USB camera with Coral Edge TPU based object detection/classification over network. The project is dependent on Google's Coral USB Accelerator.

Credits

Base code took from log0's video_streaming_with_flask_example. Coral object classification originally from Google LLC.

Usage

  1. Install the Edge TPU runtime and Python library following https://coral.withgoogle.com/docs/accelerator/get-started/
  2. Install Python dependencies: cv2, flask.
  3. Download pretrained Edge TPU sample models using: sh download_models.sh
  4. Run "python3 main.py", or run "sh ./start.sh" in the background.
  5. Navigate the browser to the http://ip-adress:5000.

Switch models

  1. Switch between detection/classification by modifying main.py imports
  2. Change the model used for detection/classification by modifying the config.py

Reference

  1. https://github.com/log0/video_streaming_with_flask_example
  2. https://coral.withgoogle.com/docs/edgetpu/api-intro

About

Object detection live streaming using Coral USB Accelerator

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published