How to seperate and track square in a grid image?

edit

I saw a youtube video https://www.youtube.com/watch?v=il75K.... It has 10x10 grid and I think it use corner detection to seperate each square in that grid. But I dont know how to track the square to number each block. Then I can use contour detection or neural network detection to detect X and O. And then draw exactly position in an image. Is there anyone has idea for this? Thank you