Raspberry PI-based camera-based instrumentation

Initially, we are building a service that detects when a conference room is in use at the Capital Factory and makes that information available to members. Because we are using images instead of motion detectors, we can do a lot more interesting things in the future.

Development starts 1/13/14.

Initial Team:

• Joshua Ellinger (CTO/Exemplar Technologies)
• Chris Brown (UT Grad Student)
• Connor Smith (UT CS Student)

AWS connection

The script uses boto, so the usual boto config locations (~/.boto, /etc/boto.cfg) are used. This also means that environment variables AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY with precedence as defined by boto.

1. pi-eye timelapse (run from this package on the Raspberry Pi) will upload a picture to S3 with ID "pi2" every 15 seconds. Option examples:

    --interval 60   # Upload every 60 seconds
    --rotation 180  # Rotate the camera output by 180° before uploading
    --id pi3        # Use id "pi3" (thus, upload to /images/pi3 instead)
   

Currently this is lazy and does not actually upload X seconds, but X seconds + how long it takes to capture and upload the image.

2. pi-eye app

   Assuming that you cloned this repo like into /www/pi-eye, and installed nginx, you should link in the config file so that you don't have to run the Bottle app as root, but still listen on :80, by having nginx back-proxy requests to the Bottle app listening on :8091 (the default). Run this: ln -sf /www/pi-eye/proxy.nginx /etc/nginx/sites-enabled/pi-eye