GithubHelp home page GithubHelp logo

robodhhb / smartpicam Goto Github PK

View Code? Open in Web Editor NEW
10.0 4.0 4.0 20.2 MB

Real time object detection with Raspberry Pi, Google Edge TPU and Python

License: MIT License

Python 100.00%
object-detection raspberry-pi neural-network tpu-acceleration python3 tensorflow-lite artificial-intelligence

smartpicam's Introduction

SmartPiCam

Real time object detection with Raspberry Pi, Google Edge TPU and Python

Willkommen bei der SmartPiCam! (see English version below)

Dieses Projekt realisiert eine smarte Kamera mit dem Raspberry Pi, die in Echtzeit mit Hilfe der Edge TPU von Google die Objekte "Apfel" und "Flasche" erkennt. Das neuronale Netz basiert auf einem von Google trainierten Modell auf Basis des COCO-Datensatzes mit der Architektur "MobileNet SSD v2". Das Inferencing findet auf der Edge TPU statt. Der Raspberry Pi kann dadurch 6 Bilder pro Sekunde anzeigen und erkannte Objekte kennzeichnen(TPU Konfiguration mit normaler Geschwindigkeit). Die Bilder haben eine Auflösung von 300 x 300 Pixel. Das "Video SmartPiCam HD" zeigt eine Aufnahme des Desktops vom Raspberry Pi in Echtzeit.

In den Ordnern "SmartPiCam" und "SmartPiCam_with_PyCoral_API" befinden sich jeweils eine Installationsanleitung und die Python Quellen. Das PyCoral-API ist das aktuelle API für die Edge TPU. Das Projekt ist ausführlich beschrieben im Artikel "Objekterkennung mit Pi-Kamera und Edge-TPU", deutsches Make-Magazin 03/2019, 20.06.2019, Seite 58-62. https://www.heise.de/select/make/2019/3/1561278575740040

Siehe auch: https://github.com/robodhhb/RoboPiCam
https://github.com/robodhhb/Smart-Modelrailway-Cam

Welcome to the SmartPiCam!

This project realizes a smart camera with the Raspberry Pi. It detects the objects "apple" and "bottle" in real time using the Edge TPU from Google. The neural network is based on a model trained at Google on basis of the COCO-Dataset with the architecture "MobileNet SSD v2". Inferencing is executed on the Edge TPU. The Raspberry Pi is therefore able to detect and mark objects on 6 pictures per second (TPU configuration with default operating frequency). The pictures have a resolution of 300 x 300 Pixel. The "Video SmartPiCam HD" shows a real time recording of the pi's desktop.

See folders "SmartPiCam" and "SmartPiCam_with_PyCoral_API" for an installation manual and the python sources. Please use the PyCoral-API implementation because it is the actual API for the Edge TPU. This project is described in the article "Objekterkennung mit Pi-Kamera und Edge-TPU", german Make Magazine 03/2019, 20.6.2019, p. 58-62. https://www.heise.de/select/make/2019/3/1561278575740040

See also: https://github.com/robodhhb/RoboPiCam
https://github.com/robodhhb/Smart-Modelrailway-Cam

smartpicam's People

Contributors

robodhhb avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.