Fish-eye calibration

hi all,

i’m trying to do the camera calibration with fish-eye camera fov:160° with distorion <14.3% following this example: VPI - Vision Programming Interface: Fisheye Distortion Correction. my goal is to measure distance between Aruco Tags on 2mx3m map using opencv. So i tryied multiple calibration with 10 and 185 images. The rms error output of the calibration is between 35 and 260, but the matrix stay always the same. Then when i use the fonction estimatePosSingleMarkers. Listing rvecs and tvecs for the distance between tags calcul. The output is correct in the center of the frame. But if one of the two tag is in the distorded zone of the image the output value is wrong.
Please someone can tell me: if i undistort the image and after estimatePosSingleMarkers the value of rvecs and tvecs won’t be represent real word right ?
So is my calibration having a problem ?
or my camera has bad lens ?
or am i trying to do something impossible ?
thanks for your time


Would you mind attaching some output images so we can know more about your issue?

More, would you mind double-checking if the output of findChessboardCorners is good or not.
It is common that the corner detection doesn’t work well due to the capturing motion blur.


This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.