Appearance Based Indexing for Relocalisation in Real-Time Visual SLAM

Denis Chekhlov; Walterio Mayol-Cuevas; Andrew Calway

doi:10.5244/C.22.37

Appearance Based Indexing for Relocalisation in Real-Time Visual SLAM

Denis Chekhlov, Walterio Mayol-Cuevas, Andrew Calway

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

29 Citations (Scopus)

Abstract

Previous work on visual SLAM has shown that indexing on space and scale facilitates the use of feature descriptors for matching in real-time systems and that this can significantly increase robustness. However, the performance gains necessarily diminish as uncertainty about camera position increases. In this paper we address this issue by introducing a further level of indexing based on appearance, using low order Haar wavelet coefficients. This enables fast look up of descriptors even when the camera is lost, hence allowing efficient relocalisation. Results of experiments on a range of real world test cases demonstrate that the method is effective, including single frame relocalisation rates up to 90\% using relatively low numbers of descriptor comparisons.

Translated title of the contribution	Appearance Based Indexing for Relocalisation in Real-Time Visual SLAM
Original language	English
Title of host publication	19th British Machine Vision Conference
Pages	363-372
Number of pages	10
DOIs	https://doi.org/10.5244/C.22.37
Publication status	Published - Sept 2008

Bibliographical note

ISBN: 9781901725360
Publisher: BMVA
Name and Venue of Conference: 19th Bristish Machine Vision Conference
Other identifier: 2000939

Access to Document

10.5244/C.22.37

http://www.cs.bris.ac.uk/Publications/pub_master.jsp?id=2000939

Handle.net

Persistent link

Cite this

@inproceedings{ead09ef9767e4ace89221df20cfcca0b,

title = "Appearance Based Indexing for Relocalisation in Real-Time Visual SLAM",

abstract = " Previous work on visual SLAM has shown that indexing on space and scale facilitates the use of feature descriptors for matching in real-time systems and that this can significantly increase robustness. However, the performance gains necessarily diminish as uncertainty about camera position increases. In this paper we address this issue by introducing a further level of indexing based on appearance, using low order Haar wavelet coefficients. This enables fast look up of descriptors even when the camera is lost, hence allowing efficient relocalisation. Results of experiments on a range of real world test cases demonstrate that the method is effective, including single frame relocalisation rates up to 90\% using relatively low numbers of descriptor comparisons. ",

author = "Denis Chekhlov and Walterio Mayol-Cuevas and Andrew Calway",

note = "ISBN: 9781901725360 Publisher: BMVA Name and Venue of Conference: 19th Bristish Machine Vision Conference Other identifier: 2000939",

year = "2008",

month = sep,

doi = "10.5244/C.22.37",

language = "English",

pages = "363--372",

booktitle = "19th British Machine Vision Conference",

}

TY - GEN

T1 - Appearance Based Indexing for Relocalisation in Real-Time Visual SLAM

AU - Chekhlov, Denis

AU - Mayol-Cuevas, Walterio

AU - Calway, Andrew

N1 - ISBN: 9781901725360 Publisher: BMVA Name and Venue of Conference: 19th Bristish Machine Vision Conference Other identifier: 2000939

PY - 2008/9

Y1 - 2008/9

N2 - Previous work on visual SLAM has shown that indexing on space and scale facilitates the use of feature descriptors for matching in real-time systems and that this can significantly increase robustness. However, the performance gains necessarily diminish as uncertainty about camera position increases. In this paper we address this issue by introducing a further level of indexing based on appearance, using low order Haar wavelet coefficients. This enables fast look up of descriptors even when the camera is lost, hence allowing efficient relocalisation. Results of experiments on a range of real world test cases demonstrate that the method is effective, including single frame relocalisation rates up to 90\% using relatively low numbers of descriptor comparisons.

AB - Previous work on visual SLAM has shown that indexing on space and scale facilitates the use of feature descriptors for matching in real-time systems and that this can significantly increase robustness. However, the performance gains necessarily diminish as uncertainty about camera position increases. In this paper we address this issue by introducing a further level of indexing based on appearance, using low order Haar wavelet coefficients. This enables fast look up of descriptors even when the camera is lost, hence allowing efficient relocalisation. Results of experiments on a range of real world test cases demonstrate that the method is effective, including single frame relocalisation rates up to 90\% using relatively low numbers of descriptor comparisons.

U2 - 10.5244/C.22.37

DO - 10.5244/C.22.37

M3 - Conference Contribution (Conference Proceeding)

SP - 363

EP - 372

BT - 19th British Machine Vision Conference

ER -

Appearance Based Indexing for Relocalisation in Real-Time Visual SLAM

Abstract

Bibliographical note

Access to Document

Handle.net

Fingerprint

Cite this