Correcting for geometrical perspective distortion

摘要

A method of correcting for geometrical perspective distortion in a sequence of digital representations of a coding pattern comprises the steps of: correcting a current digital representation for rotation in its plane, by means of a first transformation matrix; and selecting a subset of objects in the thus rotation-corrected digital representation. The method further comprises the steps of: computing a homogeneous transformation matrix based on the subset of objects and corresponding elements of said coding pattern, operating the homogeneous transformation matrix on the digital representation to produce a perspective-corrected digital representation; and updating the first transformation matrix, based on the homogeneous transformation matrix thus-computed, before processing of a subsequent digital representation.