checknext

Inspiration

What it does

Take a picture of a checkers board (from a large range of angles/tilts), and receive reccomendation on your next move, which is conveyed visually on your picture.

How we built it

OpenCV, tape, checkers board (s), pink sticky notes

Challenges we ran into

• Setting up the board detection (specifically, transforming the image to only look at the board)
• Square detection, due to the large amount of noise on the faux wood board, and some lens flares
• First time with OpenCV/computer vision
• Colors of board/pieces are very similar, so had to do some image processing

Accomplishments that we're proud of

• Inputting the checker board as a grid in a robust manner
• Piece detection, piece color detection even though color pallet of the board is monotonous

What we learned

What's next for checknext

• An actual user interface
• a better move heuristic (have seen some example checkers algorithms that use minimax that seem relatively easy to implement)
• Video streaming over HTTP from phone onto computer

Built With

• opencv-python-cameras