Reverse Vending Machine (Vegeta)

This repository contains the scripts and code for the Reverse Vending Machine project. This is a Data Science project where we used a RaspberryPI 3, and a Deep Learning model to create a reverse vending machine targeted for the tunisian market, specializing in the classification of plastic bottles and aluminum/metal cans. The users have the chance to receive points/vouchers for the type of object they insert into the vending machine.

For more information on how a reverse vending machine works :

• https://www.youtube.com/watch?v=i-AVw1gl6Xw
• https://www.youtube.com/watch?v=3dnJtUoCu2I

The "Vegeta Reverse Vending Machine" does not rely on expensive sensors to detect the type of the object inserted, it uses AI : Deep Learning to classify the object that is detected by a cheap 1080p HD IR camera.

The types of the objects which the deep learning model has been trained on :

• Plastic bottles, ranging from 25cl to 2 liters.
• Aluminum/Metal cans.

Examples

RasperryPI Setup :

A Raspberry PI 3 Model B running on Rasipan OS has been used for this project. The main code that has been used for taking the picture can be found in the Vending_Machine_Python_Script.py file.

Here's how we setup the card and the components :

For the Demo Videos :

• Reverse Vending Machine Web Application
• Reverse Vending Facial Recognition