![]() (This project is built to recognize the chessboards' images from a particular magazine, but the approach can be used for different styles of chessboards). Separation of the chessboard into 64 cells: find inside corners of the black frame and draw horizontal and verticals lines which separate the cellsĬlassification of the pieces contained in the cells: use a kNN classifier to recognize the cells(Bishop, Empty Square, King, etc.) and then find the pieces' colors and orientationsįind orientation of the chessboard: find the global orientation from the pieces orientation and eventually rotate the chessboard if the chessboard gets rotated go back to step Separation of the chessboard into 64 cells, needed for the FEN string construction and for a better analysis of the cells' contentsĬonstruction of the FEN string: build the FEN string according to the standard to describe the pieces distribution inside the chessboard Projective transformation of the image: rectify the image and crop it around the checkerboard Given an image, containing a 2D chessboard from the magazine "La settimana Enigmistica", recognize the chessboard e build the characters string according to the FEN encoding used as standard to describe the pieces distribution inside the chessboardĬhessboard segmentation: find the black square frame of the chessboard in the image
0 Comments
Leave a Reply. |