One aspect of the invention is embodied in a number of machine-readable media having stored thereon sequences of instructions which, when executed by a machine, cause the machine to perform a number of actions. The actions include, for each of a plurality of image space coordinates of a camera image, retrieving 1) a corresponding camera position, and 2) corresponding object space coordinates on an object surface that is represented by the camera image. Then, for each of a plurality of height surfaces (chosen with respect to the object surface), a grid coordinate defined by an intersection between the height surface, and a line extending between the camera position and the object space coordinates, is determined. The determined grid coordinates are then associated with their corresponding image space coordinates. Finally, coefficients for a rational polynomial camera model are estimated by fitting rational polynomials to at least the grid coordinates and image space coordinates.
展开▼