"Graz" . . "1"^^ . . "21230" . . . "2"^^ . . "414591" . "Computation of Epipolar Geometry and Trifocal Tensor from Spherical Images" . "[80FD8B6A0007]" . . "CVWW 2007: Proceedings of the 12th Computer Vision Winter Workshop" . "Computation of Epipolar Geometry and Trifocal Tensor from Spherical Images"@en . . . . "RIV/68407700:21230/07:00135417" . "Computation of Epipolar Geometry and Trifocal Tensor from Spherical Images"@en . "2007-02-06+01:00"^^ . "St. Lambrecht" . . . "RIV/68407700:21230/07:00135417!RIV11-GA0-21230___" . . . . "We formulate multiple view geometry for 3D reconstruction using a spherical-camera model as a normalized camera model expressing all central cameras. In the sense of geometry, a central camera is a collection of all rays incident to one point. Generally speaking, most cameras that are widely-used in the field of computer vision and robot vision are designed to fall into the central camera. This is because geometric and algebraic expressions of central cameras are simple compared to ones of non-central cameras. In the history of computer vision, the conventional cameras are modeled as pinhole cameras. However, the pinhole-camera model can not express catadioptric, dioptric, and panoramic cameras that also belong to the central cameras. For the use of the spherical-camera model as the normalized camera model of central cameras, we formulate the epipolar geometry and trifocal tensor for the recovery of the camera motions. The formulation based on spherical cameras simplifies the analysis" . "Omnidirectional Vision; Two and Three View Geometry"@en . "Verlag der Technischen Universit\u00E4t Graz" . . . "Torii, Akihiko" . "978-3-902465-60-3" . "Imiya, A." . "Computation of Epipolar Geometry and Trifocal Tensor from Spherical Images" . "We formulate multiple view geometry for 3D reconstruction using a spherical-camera model as a normalized camera model expressing all central cameras. In the sense of geometry, a central camera is a collection of all rays incident to one point. Generally speaking, most cameras that are widely-used in the field of computer vision and robot vision are designed to fall into the central camera. This is because geometric and algebraic expressions of central cameras are simple compared to ones of non-central cameras. In the history of computer vision, the conventional cameras are modeled as pinhole cameras. However, the pinhole-camera model can not express catadioptric, dioptric, and panoramic cameras that also belong to the central cameras. For the use of the spherical-camera model as the normalized camera model of central cameras, we formulate the epipolar geometry and trifocal tensor for the recovery of the camera motions. The formulation based on spherical cameras simplifies the analysis"@en . "P(GA201/07/1136), R" . "8"^^ .