A novel SVD based algorithm for pose estimation

In Close Range Photometry, the rapid development of low-cost digital imaging systems has led to an increased adaptation of uncalibrated consumer cameras for metrological purposes. Many well-developed techniques to derive parameters describing the unknown imaging behavior of camera and optics from a set of recorded images exist. Within an industrial measurement environment, further constraints may arise. This paper introduces a new algorithm to derive exterior camera parameters (position and orientation) from a single recorded image of a predefined test pattern. Experimental results are used to prove its applicability even in cases of partial occlusion.