Stereo Camera Calibration does not find checkerboard pattern

I am trying to use synchronized images from two cameras in order to obtain their relative rotation and translation using the Stereo Camera Calibrator App. My first step is to obtain the intrinsic matrices for each of the cameras by using the monocular Camera Calibration App, and then I use the same images as an input to the Stereo Camera Calibration. Despite the fact that I am able to calibrate each camera separately, whenever I am trying to perform the stereo camera calibration, the process of detecting the images takes very long, and after all the images are uploaded, I get the error message that no checkerboard pattern was found. I find it very strange since I am using the same images as in the monocular camera calibration!

[Gautam]

Hi Alexandru,

I'm facing the same problem, were you able to figure out what the issue was with your case?

Thanks

Answer 1

[Supraja]

The error message “no checkerboard pattern was found” usually occurs when the input image is not able to detect the checkerboard pattern and there is significant distortion or occlusion.

You can adjust the parameters of the “detectCheckerboardPoints” function i.e. “MinCornerMetric” and “MinCornerMetricThreshold”. The detailed documentation link of the function is given here:

https://www.mathworks.com/help/vision/ref/detectcheckerboardpoints.html?s_tid=doc_ta

Also try resizing the image for better resolution.

Hope this helps!