Remove lens distorsion with intrinsic or extrinsic parameters? I'm struggling with the theoretics of camera calibration. There are the intrinsic parameters and the extrinsic parameters. I know what they are but what I'm struggeling with is to understand which of these help me getting the lens distortion removed. Not even this article helps me on the question
Hope someone can clarify this?
 A: The intrinsic parameters describe the relationship between light arriving at the camera from a particular direction, and where it ends up on the sensor / focal plane. As such - once you have calibrated a camera carefully for a given focal length, those parameters are "fixed" for that camera.
But to know where something is, and how big it is etc, you ALSO need to know where the camera is "in the universe" (the coordinate system of the lab, and the relationship to the object in question). This is where extrinsic calibration comes in - it tells you "in what direction the camera is pointing", how it is rotated in space, etc.
You need to know both those things to be able to make sense of the image on the screen. Without the extrinsic calibration, you would not be able to answer the question "am I looking at the front of the object, or at the back?", or "which way is up in this image?". Without the intrinsic calibration, you don't know how to go from "this object looks like it is 1 cm tall on my screen; what is the angle between the rays coming from the top and the bottom as they arrive at the optical center of my camera?"
It follows that lens distortion is addressed with appropriate intrinsic calibration - it's a property of the camera, regardless of where it is in space.
