Coin Sorting and Detection System - VESIT Hardware Hackathon

Project Overview

This project was developed as a part of a hardware hackathon at Vivekanand Education Society's Institute of Technology (VESIT). It aims to automatically sort and detect different denominations of Indian coins using a combination of custom-designed hardware and machine learning.

We designed a CAD model for the sorting mechanism and trained a machine learning model on a custom-collected dataset of Indian coin images. The system utilizes a Raspberry Pi for image processing and control.

Features

• Automated Coin Sorting: Physical sorting of coins based on diameter.
• Coin Denomination Detection: Image-based classification of coin denominations using a trained machine learning model.
• Raspberry Pi Integration: Utilizes a Raspberry Pi for image processing and system control.
• Custom Dataset: Trained on a dataset of Indian coin images collected by the team.
• Custom CAD Design: Designed and fabricated a custom CAD model for the sorting mechanism.

Hardware Components

• Raspberry Pi
• Camera Module (for image capture)
• Sorting Mechanism (custom-designed and fabricated)
• Servo motors (for sorting control)
• Various electronic components (wires, resistors, etc.)

Software Components

• Python (for image processing and control)
• OpenCV (for image processing)
• pytorch (for machine learning model)
• Raspbian OS (on Raspberry Pi)

Hardware Setup

Coin Detection using OpenCV

Installation and Usage

1. Hardware Setup: Assemble the hardware components as per the designed CAD model and circuit diagram.
2. Software Setup:
   • Install Raspbian OS on the Raspberry Pi.
   • Install necessary Python libraries (OpenCV, etc.).
   • Copy the Python scripts and trained model to the Raspberry Pi.
3. Dataset: Place the dataset of coin images in the appropriate directory.
4. Running the System: Execute the Python script on the Raspberry Pi. The system will capture images, detect coin denominations, and control the sorting mechanism.

Project Output

Team Members

• Kartikey Pathak
• Piyush Patle
• Shaurya Rane
• Vishal Mutha

Future Enhancements

• Improve the accuracy of the machine learning model.
• Implement a user interface for easier control and monitoring.
• Add a coin counting feature.
• Optimize the hardware design for better efficiency.
• Add a display to show the detected coin values.

Acknowledgments

• Vivekanand Education Society's Institute of Technology (VESIT) for organizing the hardware hackathon.