Visual Image Processing of Humanoid Go Game Robot Based on OPENCV

Abstract

The visual system is a key part of the research on humanoid robots. Aiming at the chessboard information extraction of Go robots, a python-based chessboard recognition method is proposed. Projection transformation is used to correct the image, and binarization is used to enhance the image. Image processing techniques such as color space transformation, threshold segmentation, high-pass filtering, and Huff transform are used to detect, locate, and segment chess pieces in a chess game. Experiments show that the method can locate and segment chess pieces accurately and has a high recognition rate.