Why calibrate camera parameters in computer vision?

The purpose of camera calibration is to solve the internal and external parameters of the camera.

What parameters do you know? It depends on the role of the camera. Not all internal and external parameters are required. Maybe just to correct the lens distortion.

Camera calibration in AR is mainly to find the external parameters of the camera, that is, to determine the orientation of the camera, that is, to find the rotation matrix and translation vector.

Each camera calibration is only an approximation of the physical camera model, more specifically, each calibration is only an approximation of the physical camera model in the sampling space.

If you want to get higher accuracy in a space, you can calibrate it in layers for many times in this space, and get the imaging distance in other ways in actual calculation, so as to choose the appropriate calibration parameters.

reference data

Zhihu, Zhihu [citation time 20 18-3- 13]