Our client needed to create a dart board game where users' scores would be automatically calulated and displayed throughout the game. We were tasked with building the computer vision software used to detect dart landing locations and output their score after every throw. This project utilized 3 cameras all pointed towards the dart board that needed to reach consensus about where the dart landed.
Tools & Technologies
Python, SQL, OpenCV
Challenge
We had to ensure that homography of all 3 cameras angles were appropriately transformed and mapped to the correct placements. Dart masks were carefuly chosen so they would not be mistaken for other similar objects entering the field of view. Cailibration also proved to be a crucial aspect, which we developed a safe solution for using static board points. Lighting was an aspect that we had to control for and keep constant in the real world environment.
