Euro Coin Detection

📌 Course: BM40A1201 Digital Imaging and Image Preprocessing
📍 LUT University
👥 Contributors: Tanjuma Haque, Nada Rahali

---

📖 Overview

This project implements an imaging measurement system that detects and classifies Euro coins from images using a MATLAB-based solution.

The system performs:

• Calibration of the imaging setup (bias, dark, and flat-field corrections).
• Geometric calibration using checkerboard boxes.
• Detection and classification of coins into six denominations (€2, €1, 50c, 20c, 10c, 5c).
• Error evaluation of the detection process.

All images were provided by the professor and are not included in this repository.

---

📂 Dataset Overview

The dataset was provided via Moodle and includes:

• Bias images (sensor noise)
• Dark images (long exposure noise)
• Flat-field images (illumination correction)
• Measurement images (coins with calibration checkerboard)

Note: Due to course policy, these datasets are not included in this repository.

---

📊 Results

The system was tested on 12 measurement images.

• ✅ Average accuracy: 97.57%
• 🎯 Exact matches: 10 out of 12 images
• ⚠️ Misclassifications:
  • Case 1: One €2 coin misclassified as a 10c coin
    • Accuracy: 87.5%
    • MAE: 0.33
    • Misclassification: 12.5%
  • Case 2: One 20c coin misclassified as a 50c coin
    • Accuracy: 83.3%
    • MAE: 0.33
    • Misclassification: 16.7%

🔎 Limitations identified:

• Sensitivity to image capturing distance
• Difficulty with coins in close proximity or overlapping

✅ Overall, the system performed robustly, achieving high reliability with minor errors.

---

📂 Repository Structure

Digital-Imaging-PA/
│
├── src/                     # MATLAB source codes
│   ├── main_script.m        # Main script to run the full process
│   ├── estim_coins.m        # Function for detecting & classifying coins
│   ├── compute_mean.m       # Helper for mean calibration images
│   ├── compute_pixelToMMRatio.m   # Computes pixel-to-mm ratio
│   └── measurement_error.m  # Evaluates detection accuracy and error
│
├── docs/
│   └── PA_Digital_imaging_Final.pdf   # Final project report
│
├── README.md                # Project documentation
└── .gitignore               # Ignore MATLAB cache/temp files

---

License

This project is licensed under the MIT License – see the LICENSE file for details.