Table des matières

Fabvisio

October - December 2017
Project's holder: Romain Mekarni romain.mekarni@gmail.com

This project has been designed for the FabLab Sorbonne University. All rooms are connected by the same subnet.

Requirements

A server is required to concentrate all camera image flows. We juste use a simple computer running linux but it can be a rasberry pi to limit the energy consumption. If computers are in a router subnet (like WiFi) you need admin access to the router configuration interface.

We need one computer per video stream. Raspberry preferred.

One camera per video stream. It can be :

camera_raspberry.jpg notebook_camera.jpg

One screen per room : computer screen, TV.

Configuration

Each computer with camera runs a motion server.

This is the configuration file for simple motion server with 1 camera :

vim ~/.motion/motion.conf
output_pictures off         # motion can detect motion and save images by default. It can produce a lot of images
stream_maxrate 10           # only 10 frames per seconds is enough
stream_localhost off        # access authorized by network
stream_port 8081

This configuration file can activate a daemon service. It's a process running

  **vim /etc/systemd/system/motion.service**
[Unit]
Description=Motion daemon
After=local-fs.target network.target

[Service]
PIDFile=/var/run/motion.pid
ExecStart=/usr/bin/motion -n -c /home/fablabuser/.motion/motion.conf  <-- careful here : put your user's home folder
Restart=always
RestartSec=10
StandardOutput=syslog
StandardError=syslog
SyslogIdentifier=motion

[Install]
WantedBy=multi-user.target

Then, it becomes possible to manipulate the camera server with :

systemctl start motion
systemctl stop motion
systemctl enable motion  <-- to start motion at boot-up

One computer runs a simple http server.

Nodejs makes us easy to run a web http server with a simple command.

Blog

what are you up to

still working on raspberry Pi?

using Mac Mini can be better ?

where is the screen? entrance? (picture can be great)

how hard is to work with university network?

what tech do you think to use for making the visio

Still thinking to use Raspberry Pi cam?

Debate

You can see people who can see you. If someone really wants to monitor you, he will found a better way than a 1024×768 camera.