Plane segmentation from two views in reciprocal-polar image space

Z Z Chen; N E Pears; B J Liang; J McDermid

Plane segmentation from two views in reciprocal-polar image space

Z Z Chen, N E Pears, B J Liang, J McDermid

Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We present a new method of segmenting planar regions when an uncalibrated camera undergoes (near) pure translation. We show that, for pure translation parallel to a plane, the relation between the two views, when expressed in a reciprocal-polar (1/r, theta) space, is a pure shift in the 1/r dimension for a given value of theta. Furthermore, we show that the magnitude of these shifts follows a sinusoidal form along the theta direction over a maximum of pi radians. This allows planar image motion to be accurately detected and recovered by 1D correlation. Simultaneous planar pixel grouping, and recovery of the plane homography, thus amounts to a robustly fitting a sinusoid to shifts of maximum correlation in reciprocal-polar space. The phase of the recovered sinusoid corresponds to the orientation of the vanishing line of the plane and the amplitude is related to the magnitude of the camera translation.

Original language	English
Title of host publication	IMAGE ANALYSIS AND RECOGNITION, PT 1, PROCEEDINGS
Editors	A Campilho, M Kamel
Place of Publication	BERLIN
Publisher	Springer
Pages	638-646
Number of pages	9
ISBN (Print)	3-540-23223-0
Publication status	Published - 2004
Event	International Conference on Image Analysis and Recognition - Oporto Duration: 29 Sep 2004 → 1 Oct 2004

Conference

Conference	International Conference on Image Analysis and Recognition
City	Oporto
Period	29/09/04 → 1/10/04

Cite this

@inproceedings{695cc6f6f8fd41709c25803741265dc4,

title = "Plane segmentation from two views in reciprocal-polar image space",

abstract = "We present a new method of segmenting planar regions when an uncalibrated camera undergoes (near) pure translation. We show that, for pure translation parallel to a plane, the relation between the two views, when expressed in a reciprocal-polar (1/r, theta) space, is a pure shift in the 1/r dimension for a given value of theta. Furthermore, we show that the magnitude of these shifts follows a sinusoidal form along the theta direction over a maximum of pi radians. This allows planar image motion to be accurately detected and recovered by 1D correlation. Simultaneous planar pixel grouping, and recovery of the plane homography, thus amounts to a robustly fitting a sinusoid to shifts of maximum correlation in reciprocal-polar space. The phase of the recovered sinusoid corresponds to the orientation of the vanishing line of the plane and the amplitude is related to the magnitude of the camera translation.",

author = "Chen, {Z Z} and Pears, {N E} and Liang, {B J} and J McDermid",

year = "2004",

language = "English",

isbn = "3-540-23223-0",

pages = "638--646",

editor = "A Campilho and M Kamel",

booktitle = "IMAGE ANALYSIS AND RECOGNITION, PT 1, PROCEEDINGS",

publisher = "Springer",

address = "Germany",

note = "International Conference on Image Analysis and Recognition ; Conference date: 29-09-2004 Through 01-10-2004",

}

TY - GEN

T1 - Plane segmentation from two views in reciprocal-polar image space

AU - Chen, Z Z

AU - Pears, N E

AU - Liang, B J

AU - McDermid, J

PY - 2004

Y1 - 2004

N2 - We present a new method of segmenting planar regions when an uncalibrated camera undergoes (near) pure translation. We show that, for pure translation parallel to a plane, the relation between the two views, when expressed in a reciprocal-polar (1/r, theta) space, is a pure shift in the 1/r dimension for a given value of theta. Furthermore, we show that the magnitude of these shifts follows a sinusoidal form along the theta direction over a maximum of pi radians. This allows planar image motion to be accurately detected and recovered by 1D correlation. Simultaneous planar pixel grouping, and recovery of the plane homography, thus amounts to a robustly fitting a sinusoid to shifts of maximum correlation in reciprocal-polar space. The phase of the recovered sinusoid corresponds to the orientation of the vanishing line of the plane and the amplitude is related to the magnitude of the camera translation.

AB - We present a new method of segmenting planar regions when an uncalibrated camera undergoes (near) pure translation. We show that, for pure translation parallel to a plane, the relation between the two views, when expressed in a reciprocal-polar (1/r, theta) space, is a pure shift in the 1/r dimension for a given value of theta. Furthermore, we show that the magnitude of these shifts follows a sinusoidal form along the theta direction over a maximum of pi radians. This allows planar image motion to be accurately detected and recovered by 1D correlation. Simultaneous planar pixel grouping, and recovery of the plane homography, thus amounts to a robustly fitting a sinusoid to shifts of maximum correlation in reciprocal-polar space. The phase of the recovered sinusoid corresponds to the orientation of the vanishing line of the plane and the amplitude is related to the magnitude of the camera translation.

M3 - Conference contribution

SN - 3-540-23223-0

SP - 638

EP - 646

BT - IMAGE ANALYSIS AND RECOGNITION, PT 1, PROCEEDINGS

A2 - Campilho, A

A2 - Kamel, M

PB - Springer

CY - BERLIN

T2 - International Conference on Image Analysis and Recognition

Y2 - 29 September 2004 through 1 October 2004

ER -