SEED Lab Robot Project

Details

• Co-developed an autonomous maze-navigating robot within a 4-person team, integrating computer vision, embedded control, and I2C communication to achieve the second-fastest course completion time in the cohort.
• Engineered the computer vision subsystem using Python and OpenCV, implementing ChArUco camera calibration and ArUco marker detection to extract precise real-time distance (m) and angle (deg) metrics from 2D pixel coordinates using intrinsic camera matrices.
• Designed a multithreaded vision pipeline to asynchronously process 640x480 video feeds, applying HSV color masking and dynamically generating Regions of Interest (ROIs) to robustly identify red and green directional arrows for pathfinding logic.
• Interfaced the vision processing unit with the motor control subsystem via an I2C bus, continuously transmitting packed structural telemetry (angle, distance, and directional flags) to an Arduino while maintaining a concurrent I2C connection to update a local RGB LCD display.