COMP 498 - Project

Steve Ferreira, 27477546
Section: Winter 2018 - GG
Date: March 20, 2018

I certify that this submission is my original work and meets the Faculty's Expectations of Originality.

Prequisits

• OpenCV
• C++ 11
• CMake
• Make

Compilation instructions

This project uses the C++11 standard, and you will need a recent compiler capable of supporting these features.

This project uses CMake >=3.5 (https://cmake.org/) to generate build files for any platform. It can be installed on Linux or macOS using a package manager, or Windows using the official installer. Ensure it is added to your system PATH and accessible via command line.

Then, CMake can be used in the command line to generate build files for any environment desired.

1. cd [path to assignment]
2. mkdir build
3. cd build
4. cmake -G "your generator here" ..
   • Run cmake -G to see a list of possible generators, eg. Xcode, Unix Makefiles, versions of Visual Studio, etc.
5. Build the project using whichever build system you chose.

Objectives

1. Implement Harris Corner Detection algorithm.
2. Implement simple keypoint descriptor based on SIFT.
3. Find matching keypoint descriptors between two images.
4. Determine Homography for stitching using RANSAC.
5. Stitch images together.

Results

Matching Keypoints

Refined Inlier Matches

Final Panorama

Notes

This assignment has been tested on Linux, using OpenCV 3.4.0, CMake 3.10.2, Make 4.2.1