A review on egomotion by means of differential epipolar geometry applied to the movement of a mobile robot

Abstract

The estimation of camera egomotion is an old problem in computer vision. Since the 1980s, many approaches based on both the discrete and the di erential epipolar constraint have been proposed. The discrete case is used mainly in self-calibrated stereoscopic systems, whereas the di erential case deals with a single moving camera. This article surveys several methods for 3D motion estimation unifying the mathematics convention which are then adapted to the common case of a mobile robot moving on a plane. Experimental results are given on synthetic data covering more than 0.5 million estimations. These surveyed algorithms have been programmed and are available on the Internet.

Determining the Translational Speed of a Camera from Time-Varying Optical Flow

“…These parameters, along with the required instantaneous velocities, can be estimated on the basis of an instantaneous optical flow field comprising at least eight elements [m is estimated, recovering f andḟ proceeds by exploiting explicit formulae that involve π(C, W ) [1]. Estimation of π(C, W ) can be done by applying one of a host of methods available [8].…”

Camera motion estimation through monocular normal flow vectors

“…The matched features can describe the motion correspondence and can be used to estimate the epipolar geometry and the fundamental matrix. Finally, the camera motion parameters can be obtained by decomposing the fundamental matrix [2,5]. However, establishing accurate feature correspondences is itself a very challenging task.…”

Camera motion estimation using normal flows

“…There are two major schools in the literature. One is the displacement school, which is to track the distinct features across the image frames [1] [5]. The other one is the gradient school, which is to interpolate the full optical flows from the normal flows [3] [4] [8].…”