twox

Install HDR Homebridge camera

By taking a range of images with a range of exposure times, the Raspberry Pi can generate a High Dynamic Range images. These provide much better contrast than the automatic exposure provided by the cameras, in particurlar where the lighting conditions are stark (e.g. sunlight and shadows).

sudo apt-get update
sudo apt-get upgrade
sudo apt-get install screen vim imagemagick git ffmpeg enfuse python3-pip

Enable the camera in sudo raspi-config

Get the latest node.js install for Raspberry Pi Zero from here.

Install homebridge and the ffmepg plugin, and the ui-x component that provides the service:

sudo npm install -g homebridge 
sudo npm install -g homebridge-config-ui-x
sudo npm install -g homebridge-camera-ffmpeg

Update the .bashrc:

PATH=/opt/nodejs/bin:$PATH

Edit the .homebridge/config.json to look something like:

{
  "bridge": {
    "name": "Homebridge Pi Zero",
    "username": "CD:2D:AA:E3:CE:33",
    "port": 51826,
    "pin": "123-45-678"
  },
  "description": "Homebridge HDR Camera",
  "accessories": [
  ],
  "platforms": [
    {
      "platform": "Camera-ffmpeg",
      "cameras": [
        {
          "name": "HDR Camera",
          "videoConfig": {
            "source": "-i /home/pi/hdr_latest.jpg",
            "stillImageSource": "-i /home/pi/hdr_latest.jpg",
            "maxStreams": 2,
            "maxWidth": 1296,
            "maxHeight": 972,
            "maxFPS": 1
          }
        }
      ]
    }
  ]
}    

Install the Homebridge service:

sudo /opt/nodejs/bin/hb-service install --user pi

Dowload the script and set up the HDR capture:

cd
wget https://raw.githubusercontent.com/henrikmelin/twox/master/homebridge-hdr-camera/capture.sh
mkdir images

Load it at startup by adding this line to /etc/rc.local:

runuser -l pi -c 'sh /home/pi/capture.sh'

Go to http://<IP of pi>.local:8581 and log in using admin/admin and then add this to HomeKit using hte QR code and pin (as set in the config.json file)

Add Dropbox upload support

Since this script will generate lots of images that are overwritten, it may be good to store them somewhere to build time-lapses etc. Storing them on Dropbox also provides off-site storage if you wanna use this for monitoring.

Add a new app using the Dropbox API, and set the app to have write permissions, then generate the Access Key. Then add this to the .bashrc:

export DROPBOX_ACCESS_TOKEN="**TheAccessKey**"
export CAMERA_LOCATION="name-of-room"

And then uncomment the line in capture.sh that starts with python3 upload2dropbox.py. This requires:

pip3 install dropbox

The files will upload to the folder called picameras but that can be changed in capture.sh.