A Person Detection Security Robot That Sends You Photos

In this tutorial, you will create a desk security system with no hardware other than your laptop and the built-in webcam.

Maybe you keep a box of chocolates on your desk to snack on when you are hungry. Maybe someone is eating your chocolates when you are away. You’re not sure who, but you suspect Steve. This robot will help you catch the culprit.

When someone comes to your desk, the robot will use the vision service and the ML model service to detect a person, take their photo, and text you an alert with a photo of the person.

Text message reading “Alert There is someone at your desk beware” with a photo of a person (Steve) detected by the camera as he approaches the desk.

Hardware requirements

You need the following hardware for this tutorial:

  • Computer with a webcam
    • This tutorial uses a MacBook Pro but any computer running macOS or 64-bit Linux will work
  • Mobile phone (to receive text messages)

Software requirements

You will use the following software in this tutorial:

Configure your machine on the Viam app

Install viam-server and connect to your machine

Add a new machine in the Viam app. On the machine’s page, follow the setup instructions to install viam-server on the computer you’re using for your project. Wait until your machine has successfully connected to the Viam app.

Configure the camera component

Configure your webcam so that your machine can get the video stream from your camera:

  1. On the Viam app, navigate to your machine’s page. Check that the part status dropdown in the upper left of the page, next to your machine’s name, reads “Live”; this indicates that your machine is turned on and that its instance of viam-server is in contact with the Viam app.

  2. Click the + (Create) button next to your main part in the left-hand menu and select Component. Start typing “webcam” and select camera / webcam. Give your camera a name. This tutorial uses the name cam in all example code. Click Create.

  3. Click the video path dropdown and select the webcam you’d like to use for this project from the list of suggestions.

  4. Click Save in the top right corner of the screen to save your changes.

Test your physical camera

To test your camera, go to the CONTROL tab and click to expand your camera’s panel.

Toggle View cam to the “on” position. The video feed should display. If it doesn’t, double-check that your config is saved correctly, and check the LOGS tab for errors.

Configure your services

Now that you know the camera is properly connected to your machine, it is time to add computer vision by configuring the vision service on your machine. This tutorial uses a pre-trained Machine Learning model from the Viam Registry called EfficientDet-COCO. The model can detect a variety of things, including Persons. You can see a full list of what the model can detect in labels.txt file.

If you want to train your own model instead, follow the instructions in train a model.

  1. Configure the ML model service

    Navigate to your machine’s CONFIGURE tab.

    Click the + (Create) button next to your main part in the left-hand menu and select Service. Start typing ML model and select ML model / TFLite CPU from the builtin options.

    Enter people as the name, click Add Module, then click Create.

    In the new ML Model service panel, configure your service.

    mlmodel service panel with empty sections for Model Path, and Optional Settings such as Label Path and Number of threads.

    Select Deploy model on machine for the Deployment field. Then select the viam-labs:EfficientDet-COCO model from the Select model dropdown.

  2. Configure an mlmodel detector vision service

    Click the + (Create) button next to your main part in the left-hand menu and select Service. Start typing ML model and select vision / ML model from the builtin options.

    Enter myPeopleDetector as the name, then click Create.

    In the new vision service panel, configure your service.

    Select people from the ML Model dropdown.

    vision service panel called myPeopleDetector with filled Attributes section, mlmodel_name is “people”.

Click Save in the top right corner of the screen.

Test the detection camera

To test that the vision service is working, click on the Test area of the vision service panel myPeopleDetector or click on the vision service panel on the CONTROL tab.

the control tab

If the vision service detects a person on the configured camera, you will see a red box around the detection along with the confidence score of the detection.

How to use yagmail

Install yagmail (Yet Another Gmail/SMTP client) by running the following command in a terminal on your computer:

pip3 install yagmail

Then we have to indicate whom to send a message to, the subject, and the contents of the text message (which can be a string, image, or audio). Example code below (though you don’t have to use it yet, this will get used in the next section):

yag.send('phone_number@gatewayaddress.com', 'subject', contents)

For this tutorial, you are going to send the text to yourself. You will replace to@someone.com with your phone number and SMS gateway address. You can find yours here: Gateway Addresses for Mobile Phone Carrier Text Message. Some common ones:

  • AT&T: txt.att.net
  • T-Mobile:tmomail.net
  • Verizon Wireless: vtext.com

As an example, if you have T-Mobile your code will look like this:

yag.send('xxxxxxxxxx@tmomail.net', 'subject', contents)

This allows you to route the email to your phone as a text message.

Use the Viam Python SDK to control your security robot

Create the main script file

On your computer, navigate to the directory where you want to put the code for this project. Create a file there called chocolate_security.py. This will be the main script for the machine.

Copy the following code and paste it into chocolate_security.py:

import asyncio
import os
import yagmail

from viam.robot.client import RobotClient
from viam.components.camera import Camera
from viam.rpc.dial import Credentials, DialOptions
from viam.services.vision import VisionClient, VisModelConfig, \
    VisModelType, Detection
from viam.media.utils.pil import viam_to_pil_image

# Set environment variables. You can get them from your machine's CONNECT tab
api_key = os.getenv('API_KEY') or ''
api_key_id = os.getenv('API_KEY_ID') or ''
address = os.getenv('ADDRESS') or ''


async def connect():
    opts = RobotClient.Options.with_api_key(
      api_key=_api_key,
      api_key_id=api_key_id
    )
    return await RobotClient.at_address(address, opts)


async def main():
    machine = await connect()
    # make sure that your detector name in the app matches "myPeopleDetector"
    myPeopleDetector = VisionClient.from_robot(machine, "myPeopleDetector")
    # make sure that your camera name in the app matches "my-camera"
    my_camera = Camera.from_robot(robot=machine, name="cam")

    while True:
        img = await my_camera.get_image(mime_type="image/jpeg")
        detections = await myPeopleDetector.get_detections(img)

        found = False
        for d in detections:
            if d.confidence > 0.8 and d.class_name.lower() == "person":
                print("This is a person!")
                found = True
        if found:
            print("sending a message")
            # Convert to PIL image
            pil_image = viam_to_pil_image(img)
            # Change this path to your own
            pil_image.save('/yourpath/foundyou.jpeg')
            # Yagmail section
            # Create a yagmail.SMTP instance
            # to initialize the server connection.
            # Replace username and password with actual credentials.
            yag = yagmail.SMTP('mygmailusername', 'mygmailpassword')
            # Specify the message contents
            contents = ['There is someone at your desk - beware',
                        '/yourpath/foundyou.jpeg']
            # Add phone number and gateway address
            # found in the SMS gateway step
            yag.send('xxx-xxx-xxxx@tmomail.net', 'subject', contents)

            # If the machine detects a person and sends a text, we don't need
            # it to keep sending us more texts so we sleep it for 60
            # seconds before looking for a person again
            await asyncio.sleep(60)
        else:
            print("There's nobody here, don't send a message")
            await asyncio.sleep(10)
    await asyncio.sleep(5)
    await machine.close()

if __name__ == '__main__':
    asyncio.run(main())

After connecting to your machine, your machine will continuously check images for detections of people. If a person is detected, your robot will print sending a message, take a photo, and save it to your computer as foundyou.png (or whatever name you want).

Then it will create a yagmail.SMTP instance to send the email and send the email.

Connect the code to the robot

You need to tell the code how to access your specific machine (which in this case represents your computer and its webcam).

Navigate to the CONNECT tab on the Viam app. Make sure Python is selected in the Language selector. Get the machine address from the top of the Code Sample page and an API key and API key ID from the API keys page and set them as environment variables or add them at the top of chocolate_security.py.

To show your machine’s API key in the sample code, toggle Include API key.

Save your code file.

Run the code

You are ready to test your machine!

From a command line on your computer, navigate to the project directory and run the code with this command:

python3 chocolate_security.py

If you are in front of your computer’s webcam, you should get a text!

Your terminal should look like this as your project runs if you are in front of the camera for a bit, and then move away from the screen:

python3 chocolate_security.py
This is a person!
sending a message
x_min: 7
y_min: 0
x_max: 543
y_max: 480
confidence: 0.94140625

This is a person!
sending a message
x_min: 51
y_min: 0
x_max: 588
y_max: 480
confidence: 0.9375

This is a person!
sending a message
There's nobody here, don't send a message
There's nobody here, don't send a message

Next steps

In this tutorial, you learned how to build a security robot using the vision service, the ML model service, your computer, and your mobile phone, and we all learned not to trust Steve.

Have you heard about the chocolate box thief? He’s always got a few Twix up his Steve.

For more robotics projects, check out our other tutorials.

You can also ask questions in the Community Discord and we will be happy to help.