Multiple view geometry

This chapter introduces the concepts of projective geometry and tensor calculus and discusses how they can be applied to recover 3D models from images to yield models suited for automatic processing on real image data.

Abstract

There exist intricate geometric relations between multiple views of a 3D
\nscene. These relations are related to the camera motion and calibration as
\nwell as to the scene structure. In this chapter we introduce these concepts
\nand discuss how they can be applied to recover 3D models from images.
\nIn Section 3.2 a rather thorough description of projective geometry is
\ngiven. Section 3.3 gives a short introduction to tensor calculus and Section
\n3.4 describes in detail the camera model used. In Section 3.5 a modern
\napproach to multiple view geometry is presented and in Section 3.6 simple
\nstructure and motion algorithms are presented. In Section 3.7 more advanced
\nalgorithms are presented that are suited for automatic processing on
\nreal image data. Section 3.8 discusses the possibility of calibrating the camera
\nfrom images. Section 3.9 describes how the depth can be computed for
\nmost image pixels and Section 3.10 presents how the results of the previous
\nsections can be combined to yield 3D models, render novel views or combine
\nreal and virtual elements in video.