Automated Go Record System and Method Based on Circular Hough Transform

Illumination-Robust Automated Go Record System Based on Circular Hough Transform and Difference Imaging

This technology is a system and method that automatically records Go game moves by converting images of the entire board into grayscale and binary formats, then applying a circular Hough transform to edge detection results to identify the circular shapes and center coordinates of the stones. By performing edge detection only on the difference between the previous and current move images, it efficiently identifies newly placed stones.

Professional Go matches are still played offline, requiring manual record-keeping, while methods using intersection buttons or electrical signals require specialized boards, hindering widespread adoption. Existing image processing methods relying on intersection colors or simple image differences are vulnerable to lighting changes or slight shifts in stone positions caused by players, leading to detection errors.

This technology removes grid lines via Gaussian filtering, performs Canny edge detection and circular Hough transform on the difference image, and distinguishes between black and white stones by comparing detected center coordinates with binary pixel values. By recognizing the circular shape of the stones, it accurately tracks positions despite lighting changes or minor shifts, and reduces processing time by focusing only on changed areas. It can be used in Go clubs, professional broadcasts, and educational content using standard boards and cameras, providing a foundation for real-time game database construction without specialized equipment.

‍

‍

Key Features:

• A configuration that captures the entire Go board, converts the image to grayscale to generate a black-and-white image, and then binarizes and stores it.
• A noise reduction unit that removes grid lines using Gaussian filtering, and a difference image generation unit that creates a difference image between the previous and current stone position images.
• An automated recording unit that performs Canny edge detection on the difference image to obtain edges, then detects circles via circular Hough transform to calculate stone center coordinates.
• A step that distinguishes between black and white stones by matching center coordinates with binary pixel values, and calculates stone positions and counts to record them in the game log.

‍