Chess Board Detection

This is an algorithm for the detection of square chess boards in images regardless of various pieces on the board, which is a big limitation of cv2.findChessboardCorners().
Meant for: Python 3.8

Usage

Install requirements

pip install -r requirements.txt

Pass the following arguments to detect.py:

usage: detect.py [-h] [-f FILENAME] [-s]

Get location of chessboard on screen. Use -f to input filename (default will take a screenshot). Use -s to show result.

options:
  -h, --help            show this help message and exit
  -f FILENAME, --filename FILENAME
                        Input image filename, otherwise screenshot will be taken
  -s, --show            Show result of detection

Input filename or a screenshot of screen will be taken. Output should be something like this:

python detect.py -f ./README_Imgs/chessboard1.png -s

Region: (335, 85, 772, 771)
Pixels per Square: 96

Breakdown

Step	Visual
Board detection is achieved through the use of image manipulation with OpenCV.
Using cv2.Canny() filters the edges of the image, allowing for later use of cv2.getStructuringElement() and cv2.erode() to extract horizontal/vertical lines only.
The next step is to combine the two filtered images into one with cv2.maximum(img1, img2).
The resultant image now contains only straight and horizontal lines, allowing for contour detection around the vertices of the squares within the image.
After filtering through contours, 81 points are left. The corners are defined by the leftmost/rightmost, upmost/downmost points.