Small office doorbell system using a Slack Incoming Webhooks, triggered by motion detection using Motion, and evaluated by Google Cloud Vision API.
I am using a Raspberry Pi with a cheap USB webcam, but any system with any attached camera should be fine.
TODO: add pictures
The current setup depends on Motion, a popular program for motion detection in webcams (also does streaming).
To install Motion in Ubuntu (or Raspbian):
sudo apt-get install motion
Follow the
Motion configuration guide,
or modify motion.conf.example in this repo (which has been stripped down to bare minimum settings for clarity).
Run motion with your edited config file:
motion -c ./motion.conf.example
Follow the guide on Creating Incoming Webhooks for your Slack team.
Customize it how you like. But you need to use the URL you are given on that page.
Would be nice if this repo was transformed into a bot, so the team owner doesn't have to configure it themselves, but I don't feel like hosting this or submitting it to the app registry, etc. (maybe someday).
Currently this only supports the use of a Service Account credentials file, or if none is provided, your gcloud default credentials.
Follow the guide on Service Accounts, and make sure to enable Cloud Vision API in your project.
Use a virtualenv as usual:
virtualenv venv
source venv/bin/activate
pip install -r requirements.txt
Not an actual Slack Bot, but you can host this wherever you want and have it read from your camera.
python main.py \
--motion-output-dir /mycam \
--stream-addr myExternalIP:8081 \
--json-keyfile /path/to/service-account-credentials.json \
--webhook-url https://hooks.slack.com/services/my/webhook/url
For full list of arguments, run python main.py -h
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent