Domestic Tranqulity in the 21st century

Uses AI Vision to insure the domestic tranquility by keeping an eye on how clean your kitchen is.

Setting up

Installing python dependencies

pip3 install -r requirements.txt
sudo apt update && sudo apt install -y mpg321

Camera configuration and preview

You need to customize the framegrab.yaml file to point to the correct camera. See framegrab for reference, but this file is a pretty good starting point. It can use a local Raspberry Pi or USB camera, or most any networked RTSP camera.

You can check that the camera code is all working properly by seeing a preview in your terminal (if you have an advanced terminal program such as iTerm2).

python3 trycamera.py

Stashing your secrets

Put your secrets in the file .secret-env that looks like:

export CAMERA_PASSWORD=rtsp-password...
export GROUNDLIGHT_API_TOKEN=api_...

Then before you run this code run:

source .secret-env

Groundlight account setup

You can use a free Groundlight account. Then get an API token and save it as an environment variable:

export GROUNDLIGHT_API_TOKEN="api_..."

Running the real thing

python3 app.py

You might want to edit the motion detection parameters to make it more or less sensitive.

Hardware

I used a Raspberry Pi 4. It will work anywhere you have the right python and libraries installed. I used Groundlight's pre-built Raspberry Pi image.

Audio

On Raspberry pi, alsa is what I've used. You can check that it's working with:

speaker-test -t wav -c 2

Try editing your /etc/asound.conf file to set the correct device. Then run:

sudo alsa force-reload

And set the environment variable so pygame uses it:

export SDL_AUDIODRIVER=alsa

You might want to add this to your .secret-env file so you don't have to keep setting it up.