ORCID

Abstract

This paper intends to present and analyze all the necessary steps one needs totake in order to build a simple application for structure from motion inpractice using only the epipolar constraint. In other words, toillustrateinpractice (i.e,, cook-book style) about how the most fundamental of concepts inepipolar geometry can be combined into a complete and disambiguated seriesof steps for camera motion estimation and 3D reconstruction in a relativelysimple computer program.In particular, the essential matrix is re-introduced inorder to avoid misconceptions regarding the nature of the rotation matrix andtranslation between two camera positions during consequent computations.Strictly based on the formulation of the essential matrix in this paper, the mostimportant of its properties are given with proofs, merely to provide necessaryintuition into the series of derivations that ultimately lead to the recovery of theorientation and baseline vector (up-to-scale). Finally, an application thatperforms 3D position tracking of a simple web-camera is implemented (andsource code made available) in order to demonstrate the practical value of theessential matrix with only a few of assumptions loosely in place (stillbackground and relatively reliable point tracking).

Publication Date

2013-03-11

Volume

MIDAS.SNMSE.2013.TR.007

Publisher

MIDAS - Marine and Industrial Dynamic Analysis, Plymouth University

Deposit Date

2025-08-08

Download

Share

COinS