Jogador de xadrez robótico com visão computacional

Abstract

This project creates an automated chess player that, through computer vision and automation, can detect the opponent's moves and respond with their own moves autonomously on a physical board. To this end, a board was built, containing two perpendicular axes moved with stepper motors, which together position an electromagnet responsible for moving the chess pieces. A camera is mounted on this structure to capture images from the top of the board and, using the OpenCV library, recognize each object in the scene. For the processing of the chess move, the engine StockFish is being used as the main AI algorithm, while for the analysis a proprietary AI based on the literature on DEEP BLUE and on StockFish itself was created. The computer responsible for executing the entire application is a Raspberry Pi. [resumo fornecido pelo autor]