Counting the number of people present in our coworking space.
Post blog : http://ants.builders/blog/26-11-2014/counting-coworkers-with-machine-learning.html
In order to have some metrics about how many people use this collaborative space, we have installed a small rapsberry pi whose role is to measure affluence and pipe the result to a web page accessible by anyone.
With a constant time step, the raspberry pi takes a picture with the raspicam. This image is then processed by a computer vision algorithm that detects all the visible human heads. The number of detected people is then piped to plotly.
At the same time, we ask the network how many devices are connected and among them which one are of Apple brand.
This no spying because the only information getting out of the sensor is the number of persons and the number of devices.
sudo python stream.py -i wlan0 -s 10.185.62.0 -r "Lexmark International Inc.;NEC Corporation"
For Node :
sudo python nodeKeeper/stream.py -i wlan0 -s 10.33.0.0 -r "Lexmark International Inc.;NEC Corporation;Routerboard.com;ASUSTEK COMPUTER INC.;'Shenzhen Tp-Link Technology Co; Ltd.'"