CountMeIn

CountMeIn is a simple application that detects motion near doors and entrances with IoT devices (clients) and publishes the event to a server via a RESTful HTTP API. Use cases for CountMeIn are people counting in stores and surveillance.

The Camera Node client detects and tracks faces from a live webcam stream, then sends the face to the server as soon as it moves out the camera.

CountMeIn's server is easily extensible through plugins that are notified as soon as an event is published to the server. For instance, the Emotion Detection Plugin exploits Microsoft APIs to enrich an event generated by a camera with information about the gender, age and emotions. The Telegram Bot Plugin, instead, forwards all the events to a chat. To use these services, you must set API keys in the configuration file.

Installation and running

The easiest way to run the server is to install Docker and follow the instructions under "With Docker".

With Docker

Build and run the multi-container application with:

$ docker-compose up

Without Docker

Otherwise, you need to install OpenCV 3.2.0 with extra modules and Python 3 bindings. Also, you need to install the required Python packages:

$ pip3 install -U gunicorn hug telepot pymongo cognitive_face yapsy jinja2

Now you can run the application on a development server with:

$ python3 -m counterserver

Or using gunicorn web server:

$ pip3 install gunicorn
$ gunicorn --reload --bind=0.0.0.0:8000 counterserver.__main__:__hug_wsgi__

Testing the server is running

Try to create an event with the command:

$ curl -i http://localhost:8000/v1/events -F event_type=0

You should get a 201 Created response. You can browse http://localhost:8000/ to see the dashboard.

Now try this API endpoint to get the list of face locations in an image:

$ wget https://static.pexels.com/photos/109919/pexels-photo-109919.jpeg -O test_crowd.jpg
$ curl -i $ curl -i http://localhost:8000/v1/frame -F event_image=@test_crowd.jpg

Testing the Camera Node client

This command will start capturing from your webcam:

$ python3 clients/camera_node.py

License

This project is licensed under the terms of the GNU General Public License v3.0.

premhunt / countmein