Virtual home security doorbell system powered by facial and speech recognition.
Ryan Hebert, Daniel Kantor & Patrick Kim CSI 370
pip3 install cmake
pip3 install face_recognition
pip3 install numpy
pip3 install dlib
pip3 install opencv-python
pip3 install maria-db
pip3 install pyimgur
Navigate to the ngrok site to create an account and download the ngrok client. Run the commands below to unzip and authenticate your account with the provided auth token:
unzip /path/to/ngrok.zip
ngrok config add-authtoken <YOUR-TOKEN-HERE>
When completed, run the command below to begin the local host tunnel.
ngrok http <PORT>
Make sure the port used in the command is the same port that the Flask app runs in your code. After connected, copy the value of the forwarding key and save it for the next step.
To create a Twilio account, go to their website and create a trial account. Once granted a number, save the 10 digit (including the +1) for the from_number parameter in send message methods. Navigate to the Active Numbers tab and select your provided phone number. At the bottom of the page, enter the ngrok subdomain concatenated with "/sms" in the field that reads "A message comes in". This value is the address to which the webhook will look for to retrieve messages.
To run DoorBot, ensure a camera is connected to the machine. First, make sure ngrok is connected and running. Next, run the Flask-server.py script and wait for a success message. Finally, run Bot.py in a separate shell.
This code is meant to be run on a Raspberry Pi 4, model B. Minor complciations may arise with different machines or operating systems. The Bot has a limited understanding of natural language. If your response seems to stump the Bot, please restart. The Bot expects single text responses.