Atmosphere Station

Branch	Travis
master

The app runs on RPi and detects faces using Haar cascade classifier from video stream. Once the face detected it sends the image via HTTP to the configurable endpoint.

Running application

Make sure that all required packages installed from apt-requirments.txt and requirments.txt files. OpenCV is not available on Raspbian Jessie and needs to be compiled on RaspberryPi. Once all requirments meat simply execute:

python . -e https://posttestserver.com/post -t SECRET_TOKEN

On Raspberry Pi you will have to compile OpenCV from source.

Running as daemon service

Add environment variable to configure endpoint to /etc/environment:

ATMOSPHERE_ENDPOINT=http://...
ATMOSPHERE_TOKEN=SECRET_...

Then execute sudo visudo and add:

Defaults    env_keep +="ATMOSPHERE_ENDPOINT"
Defaults    env_keep +="ATMOSPHERE_TOKEN"

Assuming that the project cloned into /home/pi/atmosphere-station (this can me changed in atmosphere-station.sh), copy init script to /etc/init.d:

sudo cp atmosphere-station.sh /etc/init.d

Make sure that both atmosphere-station.sh and __main__.py are executable:

sudo chmod 755 /etc/init.d/atmosphere-station.sh
chmod 755 __main__.py

At this stage you can start and stop service:

sudo /etc/init.d/atmosphere-station.sh start
sudo /etc/init.d/atmosphere-station.sh status
sudo /etc/init.d/atmosphere-station.sh stop

Note that when service started it will perform git pull to update the app. See details in atmosphere-station.sh file

See log file at /tmp/atmosphere-station.log for info or errors.

Now, add it to the boot sequence:

sudo update-rc.d atmosphere-station.sh defaults