Alexa Computer Vision Face Detection Door Bell Thingy
Copy the image to your microSD card
Connect and power the pi like this (or some other way): https://learn.adafruit.com/adafruits-raspberry-pi-lesson-5-using-a-console-cable/overview
Change your hostname:
sudo /etc/hostname
sudo /etc/hostsRestart
Set up wifi by adding or updating a network
sudo vim /etc/wpa_supplicant/wpa_supplicant.confRestart
Install aws cli
sudo pip install awsclipip3 install virtualenv
pip3 install virtualenvwrapper
export WORKON_HOME=~/Envs
echo -e "\nexport VIRTUALENVWRAPPER_PYTHON=\`which python3\`" >> ~/.bash_profile
source /usr/local/bin/virtualenvwrapper.sh
mkvirtualenv cv
workon cv
pip install -r requirements.txtCamera is attached to a RasPi with a motoion detector.
Things to do!
- When motion is detected we take a picture and send it to the middle tier.
- It would be great if we could only send a picture if we have identified that there is a face in the image.
The Middle is the api layer, it is responsible for: Things to do!
- Receiving events from the camera
- Sending events to Alexa
- Maintaining the event log. This includes a log of:
- When there was someone at the door.
- Who was at the door.
- When the last time each person was at the door.
- Connecting to or running a system for facial recognition. This will likely be an external service. Google, Amazon and Algorithmia all have API's for facial recognition.
When an Alexa skill is triggered then look at the last identified face ( or event depending on the question ) and announce who it was. Depending on the question asked there should be additional information Alexa could announce, such as if there is anyone at the door or not.
- Look into Alexa Skills and Intents.
- Alexa is based on lambda functions.