This classroom-tested and easy-to-understand textbook/reference describes the state of the art in 3D reconstruction from multiple images, taking into consideration all aspects of programming and implementation. Unlike other computer vision textbooks, this guide takes a unique approach in which the initial focus is on practical application and the procedures necessary to actually build a computer vision system. The theoretical background is then briefly explained afterwards, highlighting how one can quickly and simply obtain the desired result without knowing the derivation of the mathematical detail. Features: reviews the fundamental algorithms underlying computer vision; describes the latest techniques for 3D reconstruction from multiple images; summarizes the mathematical theory behind statistical error analysis for general geometric estimation problems; presents derivations at the end of each chapter, with solutions supplied at the end of the book; provides additional material at an associated website. This classroom-tested and easy-to-understand textbook/reference describes the state of the art in 3D reconstruction from multiple images, taking into consideration all aspects of programming and implementation. Unlike other textbooks on computer vision, this Guide to 3D Vision Computation takes a unique approach in which the initial focus is on practical application and the procedures necessary to actually build a computer vision system. The theoretical background is then briefly explained afterwards, highlighting how one can quickly and simply obtain the desired result without knowing the derivation of the mathematical detail. Topics and features : Reviews the fundamental algorithms underlying computer vision, and their implementation Describes the latest techniques for 3D reconstruction from multiple images Summarizes the mathematical theory behind statistical error analysis for general geometric estimation problems Offers examples of experimental results, enabling the reader to get a feeling of what can be done using each procedure Presents derivations and justifications as problems at the end of each chapter, with solutions supplied at the end of the book Explains the historical background for each topic in the supplemental notes at the end of each chapter Provides additional material at an associated website, include sample code for typical procedures to help readers implement the algorithms described in the book This accessible work will be of great value to students on introductory computer vision courses. Serving as both as a practical programming guidebook and a useful reference on mathematics for computer vision, it is suitable for practitioners seeking to implement computer vision algorithms as well as for theoreticians wishing to know the underlying mathematical detail Front Matter....Pages i-xi Introduction....Pages 1-8 Front Matter....Pages 9-9 Ellipse Fitting....Pages 11-32 Fundamental Matrix Computation....Pages 33-57 Triangulation....Pages 59-68 3D Reconstruction from Two Views....Pages 69-80 Homography Computation....Pages 81-97 Planar Triangulation....Pages 99-105 3D Reconstruction of a Plane....Pages 107-115 Ellipse Analysis and 3D Computation of Circles....Pages 117-129 Front Matter....Pages 131-131 Multiview Triangulation....Pages 133-147 Bundle Adjustment....Pages 149-161 Self-calibration of Affine Cameras....Pages 163-182 Self-calibration of Perspective Cameras....Pages 183-210 Front Matter....Pages 211-211 Accuracy of Geometric Estimation....Pages 213-229 Maximum Likelihood of Geometric Estimation....Pages 231-242 Theoretical Accuracy Limit....Pages 243-254 Back Matter....Pages 255-321
