Space Variant Feature Extraction for Omni-directional Images

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In recent years, the use of omni-directional cameras has become increasingly morepopular in vision systems and robotics. To date, most of the research relating to omni-directional cameras has focussed on the design of the camera or the way in which toproject the omni-directional image to a panoramic view rather than on how to processthese images after capture. Typically images obtained from omni-directional cameras are transformed to sparse panoramic images that are interpolated to obtain a complete panoramic view prior to low level image processing. This interpolation presents a significant computational overhead with respect to real-time vision. We present an approach to real-time vision that projects an omni-directional image to a sparse panoramic image and directly processes this sparse image. Feature extraction operators previously designed by the authors are used in this approach but this paper highlights the reduction of the computational overheads of processing images arising from omni-directional camerasthrough efficient coding and storage, whilst retaining accuracy sufficient for application to real-time robot vision.
LanguageEnglish
Title of host publicationUnknown Host Publication
Pages44-51
Number of pages8
Publication statusPublished - 2006
EventIrish Machine Vision and Image Processing Conference 2006 - Dublin City University
Duration: 1 Jan 2006 → …

Conference

ConferenceIrish Machine Vision and Image Processing Conference 2006
Period1/01/06 → …

Fingerprint

Feature extraction
Cameras
Image processing
Computer vision
Interpolation
Robotics

Keywords

  • Omni-directional imaging
  • Feature detection

Cite this

@inproceedings{6d790fb4fa334fa4934fae4c3d055462,
title = "Space Variant Feature Extraction for Omni-directional Images",
abstract = "In recent years, the use of omni-directional cameras has become increasingly morepopular in vision systems and robotics. To date, most of the research relating to omni-directional cameras has focussed on the design of the camera or the way in which toproject the omni-directional image to a panoramic view rather than on how to processthese images after capture. Typically images obtained from omni-directional cameras are transformed to sparse panoramic images that are interpolated to obtain a complete panoramic view prior to low level image processing. This interpolation presents a significant computational overhead with respect to real-time vision. We present an approach to real-time vision that projects an omni-directional image to a sparse panoramic image and directly processes this sparse image. Feature extraction operators previously designed by the authors are used in this approach but this paper highlights the reduction of the computational overheads of processing images arising from omni-directional camerasthrough efficient coding and storage, whilst retaining accuracy sufficient for application to real-time robot vision.",
keywords = "Omni-directional imaging, Feature detection",
author = "D Kerr and BW Scotney and SA Coleman",
year = "2006",
language = "English",
isbn = "0-9553885-0-3",
pages = "44--51",
booktitle = "Unknown Host Publication",

}

Kerr, D, Scotney, BW & Coleman, SA 2006, Space Variant Feature Extraction for Omni-directional Images. in Unknown Host Publication. pp. 44-51, Irish Machine Vision and Image Processing Conference 2006, 1/01/06.

Space Variant Feature Extraction for Omni-directional Images. / Kerr, D; Scotney, BW; Coleman, SA.

Unknown Host Publication. 2006. p. 44-51.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Space Variant Feature Extraction for Omni-directional Images

AU - Kerr, D

AU - Scotney, BW

AU - Coleman, SA

PY - 2006

Y1 - 2006

N2 - In recent years, the use of omni-directional cameras has become increasingly morepopular in vision systems and robotics. To date, most of the research relating to omni-directional cameras has focussed on the design of the camera or the way in which toproject the omni-directional image to a panoramic view rather than on how to processthese images after capture. Typically images obtained from omni-directional cameras are transformed to sparse panoramic images that are interpolated to obtain a complete panoramic view prior to low level image processing. This interpolation presents a significant computational overhead with respect to real-time vision. We present an approach to real-time vision that projects an omni-directional image to a sparse panoramic image and directly processes this sparse image. Feature extraction operators previously designed by the authors are used in this approach but this paper highlights the reduction of the computational overheads of processing images arising from omni-directional camerasthrough efficient coding and storage, whilst retaining accuracy sufficient for application to real-time robot vision.

AB - In recent years, the use of omni-directional cameras has become increasingly morepopular in vision systems and robotics. To date, most of the research relating to omni-directional cameras has focussed on the design of the camera or the way in which toproject the omni-directional image to a panoramic view rather than on how to processthese images after capture. Typically images obtained from omni-directional cameras are transformed to sparse panoramic images that are interpolated to obtain a complete panoramic view prior to low level image processing. This interpolation presents a significant computational overhead with respect to real-time vision. We present an approach to real-time vision that projects an omni-directional image to a sparse panoramic image and directly processes this sparse image. Feature extraction operators previously designed by the authors are used in this approach but this paper highlights the reduction of the computational overheads of processing images arising from omni-directional camerasthrough efficient coding and storage, whilst retaining accuracy sufficient for application to real-time robot vision.

KW - Omni-directional imaging

KW - Feature detection

M3 - Conference contribution

SN - 0-9553885-0-3

SP - 44

EP - 51

BT - Unknown Host Publication

ER -