Chess piece movement detection and tracking, a vision system framework for autonomous chess playing robot

Abstract

Chess-playing robot is robots that is able to see, think, and move the chess piece properly. We separated the work into three parts (mechanics, engine, and vision), in order to ensure each of the parts work robustly. This paper is one of the series of those parts and will only focus on vision system for autonomous chess playing robot. Chessboard and chess piece detection is not a trivial task due to material and design diversity. A good chess-playing robot is able to see the chessboard and chess piece existence, and able to track chess piece movement generally and accurately. We propose a framework of chess piece movement detection algorithm. In general, there are chessboard detection, chess cell detection, chess piece movement detection, and notation creation. Contour Tracing is performed to select the region of interest, the chessboard, automatically, then Hough Transformation performed to identify the cells, then chess piece movement determination is perform by separating the Movement Detection into Stable and Unstable Frames Detection and Chess Piece Movement Detection. Finally the movement is converted into standard output, Portable Game Notation.