Three Dimensional Computer Vision by Olivier Faugeras

A Geometric Viewpoint

A rigorous, mathematically grounded treatment of recovering three-dimensional structure and motion from images, covering camera models, projective and affine geometry, epipolar geometry, stereo correspondence, motion analysis, and reconstruction algorithms. It develops theoretical foundations—multiview geometry, calibration, invariants—and presents practical methods for stereo vision, structure-from-motion, and camera calibration, emphasizing algorithm derivation, proofs, and applications to photogrammetry and robotics. The presentation balances formal mathematics with algorithmic techniques to support both theoretical understanding and implementation of 3D computer vision systems.