ArUco Robot Localization

Or watch on YouTube.

A camera-based localization system in C++ that estimates the robot’s position to within 4 cm on a 12×8 meter playing field. It detects ArUco tags with OpenCV and feeds the resulting estimates into the robot’s ROS2 stack. Webcam runs recorded with ros2 bag meant the pipeline could be replayed and tuned without the robot, and the whole environment ran in Docker so it behaved the same on every machine.