Feature Matching and Stitching Images

This repository contains four Python scripts demonstrating key computer vision techniques using OpenCV:

• SIFTDemo.py: Detects and matches SIFT features between two images.
• BFMDemo.py: Detects and matches ORB features using a brute-force matcher.
• FM+Homography.py: Performs feature matching with SIFT and estimates a homography to align images.
• Stitcher.py: Uses OpenCV’s Stitcher API to stitch multiple images into a panorama.

---

Table of Contents

1. Requirements
2. Installation
3. Usage
   • SIFTDemo.py
   • BFMDemo.py
   • FM+Homography.py
   • Stitcher.py
4. Notes

---

Requirements

• Python 3.6 or higher
• OpenCV
• NumPy
• Matplotlib

---

Installation

Install the dependencies via pip:

pip install opencv-python numpy matplotlib

---

Usage

SIFTDemo.py

SIFT feature detection and matching using a K-Nearest Neighbors matcher and ratio test.

python SIFTDemo.py

BFMDemo.py

ORB feature detection and brute-force matching.

python BFMDemo.py

FM+Homography.py

Matches SIFT features, applies ratio test, computes a homography with RANSAC, and shows only inlier matches.

python FM+Homography.py

Stitcher.py

Stitches multiple images into a panorama using OpenCV’s high-level Stitcher API.

python Stitcher.py --images img1.jpg img2.jpg [img3.jpg ...] --output panorama.jpg

Options:

• --images: List of input image file paths (space-separated).
• --output: Path to save the stitched panorama.
• Use -h or --help for full usage details.

Notes

• Ensure images have sufficient overlap for reliable feature matching and stitching.
• Adjust MIN_MATCH_COUNT in FM+Homography.py based on image content.
• Stitcher.py prints an error code if stitching fails; refer to OpenCV docs for details. Feel free to tweak parameters for better results.