Distortion parameters and the calculation methods

Hi, I am currently working with Reality Capture to obtain camera calibration values, specifically the intrinsic parameters and distortion coefficients. However, I have encountered some discrepancies when applying these distortion values in my calculations.

In my workflow, I project a 3D point onto the image plane, compute the undistorted pixel coordinates, and then reapply the distortion to obtain the pixel coordinates in the raw image. While my calculations have been verified using simulation software like Blender, the results do not align when using the distortion values obtained from Reality Capture.

I have reviewed the documentation and was not able to find any calculation process for distortion in Reality Capture. Could someone please provide information on the camera model and the distortion calculation process used in Reality Capture for the parameters provided by the software?

Additionally, I see that the calibration results given by RC (distortion coefficients) vary significantly each time I do the alignment with the same images and calibration settings. Did anyone experience this before?

Thanks for your time!

Hi @Hemanth0806
The mentioned process can be find here: https://dev.epicgames.com/community/learning/knowledge-base/vzwB/capturing-reality-realitycapture-xmp-camera-math