Automated Maze Solver

Overview

Used computer vision to detect a maze, compute a shortest path with BFS, and steer a physical marble maze via Arduino-controlled servos.

A camera feed is processed to extract the maze layout, then a breadth-first search determines a path while PID control translates it into tilt commands sent over serial to the Arduino.

What I learned

• • Perspective correction and maze extraction from a live camera feed
• • Breadth-first search pathfinding for shortest routes
• • Serial communication between Python and Arduino
• • PID tuning for stable marble control
• • Robust detection via thresholding and morphological operations