Authors: Shivansh Madan, Eric Ouano
A comprehensive, hands-on tutorial series covering essential computer vision tools and techniques for aerial robotics, from object detection and tracking to full Visual Odometry.
This repository provides a series of tutorials designed to bridge the gap between computer vision theory and practical application in aerial robotics. Whether you're a student, a hobbyist, or a researcher, these guides will walk you through the fundamental building blocks of spatial perception for drones and other autonomous systems. We cover foundational skills like object detection before progressing to building a complete, Vision-based navigation pipeline from scratch.
- Object Detection: Learn to use pre-trained deep learning models like YOLO to locate and classify objects in real-time.
- Object Tracking: Implement classic and modern object trackers (CSRT, KCF) to follow a specific object across multiple video frames.
- Feature Detection & Matching: Understand and implement algorithms like ORB to find and match keypoints between images.
- Motion Estimation: Learn how to calculate the camera's rotation and translation using the Essential Matrix and RANSAC.
- Visual Odometry (VO): Build a complete, step-by-step VO pipeline to track the camera's trajectory through a sequence of images.
- Data Visualization: Create dynamic, real-time visualizations of your algorithm's output using Matplotlib and OpenCV, complete with on-screen annotations.
- SLAM Fundamentals: Gain a clear understanding of the difference between Visual Odometry and a full SLAM system.