Parameterising internal camera geometry with focusing distance

A study on the variation of internal camera geometry (principal distance, principal point position and lens distortion parameters) with different focus distances has been conducted. Results demonstrate that variations of parameters are continuous and predictable, allowing a new way to describe internal camera geometry. The classical constant parameters, c, x p , y p , K 1 , K 2 , P 1 and P 2 , are replaced by continuous functions, c(γ), x p (γ), y p (γ), K 1 (γ), K 2 (γ), P 1 (γ) and P 2 (γ), where γ is a variable describing the focus position. Incorporation of γ as a metadata tag (for example, Exif header) of a photograph jointly with a parameterised definition of camera geometry would allow full use of the autofocus camera function; enabling maximum effective depth of field, better match of the plane of focus with the object’s position and higher reliability. Additionally, conducted tests suggest the parameterised definition of internal geometry could help to locate and correct linear dependences between adjusted parameters, potentially improving the precision and accuracy of calibration.