💎 "Marker-less Augmented Reality" with OpenCV and OpenGL.

C++ implementation of Lucas-Kanade-Image-Alignment

Implementation of Panorama stitching using ORB feature matching, and RANSAC for homography estimation, and stitching using Image pyramids.

Change a 2D image into 3D (Tour-Into-the-Picture)

The Random Cluster Model for Robust Geometric Fitting

Screen to world coordinates conversion using Perspective-N-Points strategy - OpenCV4 solvePnP().

Use OpenCV 2D Features framework to transform images of one scene into the same coordinate system and make a panorama.

Rotate 2D images in 3D space

MATLAB implementation of RANSAC for determining Homography Transformation Matrix for Image Stitching

Detection of a plane in 3D space using aruco markers. Based on Automatic Workspace Calibration Using Homography for Pick and Place - IEEE CASE 2023, Auckland, New Zealand | DOI: 10.1109/CASE56687.2023.10260601

Alternative to the built in homography-node in vvvv gamma. uses: https://github.com/nenoNaninu/HomographySharp. outputs correct values in much more cases than the original. can work on more than 4 points.

Simple solutions for computer vision from academic tasks

GeoCamPal is a modular image analysis toolkit with a GUI and an open-source Python environment. It integrates specialised tools for feature identification using HSV masking and manual identification, training data (COCO json) preparation, georeferencing images using a 2D homography matrix, DEM generation and further analysis.

Image stitching from scratch with almost implemented with numpy, functions from OpenCV are minimally used.

implementation of ransac (homography calculate fuction) on pytorch for gpu support

Calculates the 3x3 homography matrix from a given set of corresponding point pairs to transform between two points of view.

Stitching RGB/RGBD images into a single image.

Simple camera-to-world (z=0) coordinates transformation via homography matrix, OpenCV, C++. Uses the output from https://github.com/rodolfoap/points-picker