Abstract

Robust local image features have been used successfully in robot localization and camera pose estimation; region tracking using affine warps is considered state of the art also for many years. Although such correspondences provide a warp of the local image region and are quite powerful, in direct pose estimation they are so far only consid- ered as points and therefore three of them are required to construct a camera pose. In this contribution we show how it is possible to directly compute a pose based upon one such feature, given the plane in space where it lies. This differential correspondence concept exploits the tex- ture warp and has recently gained attention in estimation of conjugate rotations. The approach can also be considered as the limiting case of the well-known spatial resection problem when the three 3D points ap- proach each other infinitesimally close. We show that the differential correspondence is more powerful than conic correspondences while its exploitation requires nothing more complicated than the roots of a third order polynomial. We give a detailed sensitivity analysis, a comparison against state-of-the-art pose estimators and demonstrate real-world ap- plicability of the algorithm based on automatic region recognition.