Sensor Pose Estimation from Multi-center Cylindrical Panoramas

Cylindrical panoramas can be classified into various types according to their basic scanning properties and mutual spatial alignment, such as single-center (e.g., as in QTVR), concentric, multi-center, symmetric, or (after a transformation onto a cylinder) catadioptric panoramas. This paper deals with a solution of the sensor pose estimation problem using (somehow calculated) corresponding points in the multi-center panoramas. All other types of panoramas are able to be described by this general multi-center model. Due to the non-linearity of the multi-centered projection geometry, the modeling of sensor pose estimation typically results into non-linear and highly complicated forms which incur numerical instability. This paper shows that there exist linear models for sensor pose estimation under minor geometrical constraints, namely symmetric and leveled panoramas. The presented approaches are important for solving the 3D data fusion problem for multiple panoramas; it is also fundamental for an in-depth analysis of multi-view geometry of panoramic images.