Flexible Camera Calibration by Viewing a Plane from Unknown.pdfVIP

Flexible Camera Calibration By Viewing a Plane From Unknown Orientations Zhengyou Zhang Microsoft Research, One Microsoft Way, Redmond, WA 98052-6399, USA zhang@ /˜zhang Abstract Therefore, if images are taken by the same camera with fixed internal parameters, correspondences between three images We propose a flexible new technique to easily calibrate are sufficient to recover both the internal and external pa- a camera. It only requires the camera to observe a planar rameters which allow us to reconstruct 3-D structure up to pattern shown at a few (at least two) different orientations. a similarity [14, 12]. While this approach is very flexible, it Either the camera or the planar pattern can be freely moved. is not yet mature [1]. Because there are many parameters to The motion need not be known. Radial lens distortion is estimate, we cannot always obtain reliable results. modeled. The proposed procedure consists of a closed-form Other techniques exist: vanishing points for orthogonal solution, followed by a nonlinear refinement based on the directions [3, 13], and calibration from pure rotation [10, 18]. maximum likelihood criterion. Both computer simulation Our current research is focused on a desktop vision sys- and real data have been used to test the proposed technique, tem (DVS) since the potential for using DVSs is large. Cam- and very good results have been obtained. Compared with eras are becoming cheap and ubiquitous. A DVS aims at the classical techniques which use expensive equipment such as