0

I'm new to stereo vision, i'm trying to get a good stereo calibration. I calibrate each camera and i got a good results: RMS1:0.22 ,RMS2: 0.26 For the stereo calibration i got: RMS: 0.75 Is this an acceptable value? After that i used cv2.initUndistortRectifyMap and cv2.remap i got frames like this image.left frame Is this acceptable? Should i crop the image before using it to get disparity map?? thanks in advance

Imen Jmal
  • 23
  • 7

1 Answers1

0

First of all getting 0,75 RMS is not always better than getting 0,25 RMS. let me put it this way , each image have different rms value for example you took an image for calibration and its RMS is 0.10 if you take the same image making no movement on calibration plate the result RMS would be about 0.10 but your calibration wont be good. Means you cant get better undistortion.

To clarify second question you mustn't crop the images. There are a lot of reason to not do that. One of them if you crop the images you affect principal point means you apply wrong intrinsic data.

döne sarıçam
  • 1
  • 2