COMPUTER VISION-BASED METHOD AND SYSTEM FOR RECOGNIZING BLACK PIECE AND WHITE PIECE

摘要

Disclosed in embodiments of the present invention are a computer vision-based method and system for recognizing a black piece and a white piece. The method comprises: obtaining an image to be recognized on a chessboard in an actual light condition; arranging a first detection region and a second detection region on the image to be recognized; computing a first region average gray scale value in the first detection region and a second region average gray scale value in the second detection region; determining a difference and a sum between the second region average gray scale value and a preset determining threshold; comparing the first region average gray scale value with the difference and the sum, so as to determine a chess piece movement state in the first detection region. In this way, in the method, the determining threshold is set according an actual environment, so that the method is applicable to different light conditions, thereby reducing faulty detections caused by a light factor, and improving the accuracy rate for recognizing a black piece and a white piece.