Previous posts have partially answered this question, but I’m hoping to get explicit clarification of the convention in use in RC:
Principal points in the application are reported in mm. I think these are as projected on an image plane with the eq. 35mm focal length lens also reported by RC and given by: (Diagonal distance of image area in the 35mm camera (43.27mm) / Diagonal distance of image area on the image sensor of the DSC) × focal length of the lens of the DSC. So, for a sensor with an actual diagonal dimension of D, the actual focal length and principal point locations are given as f = f_35 *D/43.27. And also pp = pp_35*D/43.27. where “_35” denotes the 35mm film equivalent value.
For a Sony A6300 camera (28.2mm diagonal), and a principal point location of 0.153mm, the actual location of the pp is 0.100mm.