Published © CC BY-NC-SA

The AI Label Maker

Next level home organization is... using AI to label all your stuff?!

IntermediateFull instructions provided20 hours7,688
The AI Label Maker

Things used in this project

Hardware components

Raspberry Pi 3B+ Starter Kit
×1
Raspberry Pi Camera
×1
Adafruit Mini Thermal Receipt Printer
×1
Wall Mount AC Adapter 5V 2A
×1
Power Plug Terminal Adapter
×1
Arcade Button
×1
Thermal Sticker Paper Rolls
×1
Jumper wires
×1

Hand tools and fabrication machines

Hot glue gun
Flamingo Duct Tape
Utility knife
Metal Ruler
Scissors
Giant needle
As an Amazon Associate I earn from qualifying purchases you make using my affiliate links.

Story

Read more

Schematics

Wiring Diagram

Box Diagram

Code

LabelMakerClean.py

Python
# AI Label Maker

# Import libraries
from picamera import PiCamera
from time import sleep
import datetime
import requests
import json
import board
import busio
import adafruit_thermal_printer
import serial
import RPi.GPIO as GPIO
from time import sleep

# Setup the arcade button
GPIO.setwarnings(False)
GPIO.setmode(GPIO.BCM)
# The button GPIO pin
GPIO.setup(26, GPIO.IN, pull_up_down=GPIO.PUD_UP)
# The arcade button's LED GPIO pin
GPIO.setup(16, GPIO.OUT)

# Setup the Adafruit Mini Thermal Receipt Printer
ThermalPrinter = adafruit_thermal_printer.get_printer_class(2.68)
RX = board.RX
TX = board.TX

# For a computer, use the pyserial library for uart access
uart = serial.Serial("/dev/serial0", baudrate=19200, timeout=3000)

# Create the printer instance.
printer = ThermalPrinter(uart, auto_warm_up=True)
camera = PiCamera()

# Function that does all of the label making
def printImageCaption():
    # Set the filename of the picture to the current date and time to give it a unique name
    filename = 'image' + str(datetime.datetime.now().date()) + '_' + str(datetime.datetime.now().time()).replace(':', '.') + '.jpg'

    # Start the camera and sleep for a second to let it settle
    camera.start_preview()
    sleep(1)
    print("taking a picture")
    # Take a picture and give it the unique filename
    camera.capture('/home/pi/Desktop/labelmaker/%s' % filename)
    # Stop the camera
    camera.stop_preview()

    # Set the path to the image location
    imagepath = '/home/pi/Desktop/labelmaker/' + filename
    print("Fetching AI Label")

    # Send a request to DeepAI densecap
    # With the image and API key
    r = requests.post(
        "https://api.deepai.org/api/densecap",
        files={
            'image': open(imagepath, 'rb'),
        },
        headers={'api-key': 'YOUR-API-KEY-HERE'}
    )

    # An array of all the captions
    allcaptions = r.json().get("output").get("captions")

    # The first caption = the label the AI is the most certain of
    firstcaption = allcaptions[0]['caption']

    print(firstcaption)

    # Print the first caption using the Adafruit Mini Thermal Receipt Printer
    print("Printing Label")
    printer.justify = adafruit_thermal_printer.JUSTIFY_CENTER
    printer.print(firstcaption)
    printer.feed(6)

# Run this bit of code forever
while True:
    # If the button has been pressed
    if GPIO.input(26) == GPIO.LOW:
        # Turn on the button's LED
        GPIO.output(16, True)
        print("Button was pushed!")
        # Run the function to take a picture and print the label
        printImageCaption()
    else:
        # Turn off the button's LED
        GPIO.output(16, False)

Credits

Comments