Does someone have tips on how to improve chessboard recognition in the camera calibration?

I am currently trying to get a set of two cameras to record a timelapse and track points in it. However, I can't even get past the stereo camera calibration, as the app won't find the chessboard. Looking at the single camera calibration, at most, partial chessboards, sometimes with no recognizeable pattern behind the choice of points, are detected.

Does anybody have any suggestions on how to get the app to better recognize the chess board?

Answer 1

[Ayush]

Hi,

As per my understanding, you are facing an issue in detecting the checker pattern of a chessboard for the camera calibration. You can try using the "detectCheckerboardPoints" function to address the problem of recognizing a chessboard for camera calibration, including stereo camera setups. The function will return the detected points and dimensions of the checkerboard. A few things you should keep in mind are to ensure your chessboard has a clear, high-contrast pattern. The standard is black and white squares. Capture multiple images of the chessboard from various angles and distances. Ensure the entire chessboard is visible in each image. The general syntax for the function is:

[imagePoints,boardSize] = detectCheckerboardPoints(images)

"images" can be a single image, or an array of image filenames. "imagePoints" are the coordinates of the detected corners, and "boardSize" is the size of the chessboard.

Refer to the below documentation for more information on the "detectCheckerboardPoints" function:

https://www.mathworks.com/help/vision/ref/detectcheckerboardpoints.html