Traditional Image Processing Tutorials

This repository documents my learning process in image processing, featuring beginner-to-intermediate level exercises implemented in Python using NumPy, SciPy, and Matplotlib. Each script reflects a core concept I explored through hands-on experimentation.

📂 Contents

🟢 1. Counting Peas (counting_peas.py)

Detects and counts small round objects (e.g. peas) from a desk image using image preprocessing, contour detection, and morphological operations.

Techniques used:

• Grayscale conversion
• Thresholding
• Morphological opening
• Contour detection and counting

---

🌀 2. Image Transforms (dft_transform.py)

Applies a manual 2D Discrete Fourier Transform (DFT) and inverse DFT to visualise frequency components and reconstruct grayscale images.

Techniques used:

• 2D basis image construction
• Forward and inverse DFT (with NumPy)
• Frequency domain visualisation
• Reconstruction with RMSE comparison

---

➕ 3. Convolution Filters (convolution_filters.py)

Implements basic spatial convolution filters to detect horizontal and vertical gradients in a simple binary image.

Techniques used:

• Custom filter kernel design
• 2D convolution with scipy.signal.convolve2d
• Edge detection using gradient kernels

---

🧠 About This Project

These scripts were created as part of my self-guided learning in traditional image processing. Each file is focused and minimal to help reinforce core concepts through code.