Photogrammetric Computer Vision

Más información sobre el libro

This textbook presents a statistical perspective on the geometry of multiple view analysis, essential for camera calibration, orientation, and geometric scene reconstruction using image features. Authored by experts in geodesy and computer vision, it uniquely merges photogrammetry and computer vision. The first part introduces estimation theory, focusing on Bayesian estimation, variance components, and sequential estimation, emphasizing the importance of statistically sound diagnostics in vision metrology. The second part explores 2D and 3D geometric reasoning through projective geometry, including oriented projective geometry and statistically optimal estimation tools for geometric entities and transformations, relevant for uncertain reasoning in point clouds. The third part focuses on modeling the geometry of single and multiple cameras, covering calibration, orientation, and the statistical evaluation of corresponding scene features based on geometric image features. The authors provide algorithms for various geometric computation challenges in vision metrology, supported by mathematical justifications and statistical analyses for thorough evaluations. Each chapter is self-contained, featuring numerous figures and exercises, along with an appendix detailing basic mathematical notation and a comprehensive index. This resource is suitable for undergraduate and graduate courses in photogrammetry, computer vision, and computer