Towards accurate camera geopositioning by image matching

Raffaele Imbriaco; Clint Sebastian; Egor Bondarev; Peter H.N. De With

doi:10.1117/12.2522999

Towards accurate camera geopositioning by image matching

Raffaele Imbriaco, Clint Sebastian, Egor Bondarev, Peter H.N. De With

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

8 Downloads (Pure)

Abstract

In this work, we present a camera geopositioning system based on matching a query image against a database with panoramic images. For matching, our system uses memory vectors aggregated from global image descriptors based on convolutional features to facilitate fast searching in the database. To speed up searching, a clustering algorithm is used to balance geographical positioning and computation time. We refine the obtained position from the query image using a new outlier removal algorithm. The matching of the query image is obtained with a recall@5 larger than 90% for panorama-to-panorama matching. We cluster available panoramas from geographically adjacent locations into a single compact representation and observe computational gains of approximately 50% at the cost of only a small (approximately 3%) recall loss. Finally, we present a coordinate estimation algorithm that reduces the median geopositioning error by up to 20%.

Original language	English
Title of host publication	Eleventh International Conference on Machine Vision, ICMV 2018
Editors	Antanas Verikas, Dmitry P. Nikolaev, Petia Radeva, Jianhong Zhou
Place of Publication	Bellingham
Publisher	SPIE
Number of pages	8
ISBN (Electronic)	9781510627482
DOIs	https://doi.org/10.1117/12.2522999
Publication status	Published - 1 Jan 2019
Event	11th International Conference on Machine Vision, ICMV 2018 - Munich, Germany Duration: 1 Nov 2018 → 3 Nov 2018 Conference number: 11

Publication series

Name	Proceedings of SPIE
Volume	11041
ISSN (Print)	0277-786X

Conference

Conference	11th International Conference on Machine Vision, ICMV 2018
Abbreviated title	ICMV
Country/Territory	Germany
City	Munich
Period	1/11/18 → 3/11/18

Keywords

convolutional neural networks
geopositioning
Image matching
panoramic images
visual place recognition

Access to Document

10.1117/12.2522999

Cite this

@inproceedings{72203ab19050437790d7cdb66c2ee1ef,

title = "Towards accurate camera geopositioning by image matching",

abstract = "In this work, we present a camera geopositioning system based on matching a query image against a database with panoramic images. For matching, our system uses memory vectors aggregated from global image descriptors based on convolutional features to facilitate fast searching in the database. To speed up searching, a clustering algorithm is used to balance geographical positioning and computation time. We refine the obtained position from the query image using a new outlier removal algorithm. The matching of the query image is obtained with a recall@5 larger than 90% for panorama-to-panorama matching. We cluster available panoramas from geographically adjacent locations into a single compact representation and observe computational gains of approximately 50% at the cost of only a small (approximately 3%) recall loss. Finally, we present a coordinate estimation algorithm that reduces the median geopositioning error by up to 20%.",

keywords = "convolutional neural networks, geopositioning, Image matching, panoramic images, visual place recognition",

author = "Raffaele Imbriaco and Clint Sebastian and Egor Bondarev and {De With}, {Peter H.N.}",

year = "2019",

month = jan,

day = "1",

doi = "10.1117/12.2522999",

language = "English",

series = "Proceedings of SPIE",

publisher = "SPIE",

editor = "Antanas Verikas and Nikolaev, {Dmitry P.} and Petia Radeva and Jianhong Zhou",

booktitle = "Eleventh International Conference on Machine Vision, ICMV 2018",

address = "United States",

note = "11th International Conference on Machine Vision, ICMV 2018, ICMV ; Conference date: 01-11-2018 Through 03-11-2018",

}

Imbriaco, R, Sebastian, C, Bondarev, E & De With, PHN 2019, Towards accurate camera geopositioning by image matching. in A Verikas, DP Nikolaev, P Radeva & J Zhou (eds), Eleventh International Conference on Machine Vision, ICMV 2018., 110411C, Proceedings of SPIE, vol. 11041, SPIE, Bellingham, 11th International Conference on Machine Vision, ICMV 2018, Munich, Germany, 1/11/18. https://doi.org/10.1117/12.2522999

Towards accurate camera geopositioning by image matching. / Imbriaco, Raffaele; Sebastian, Clint; Bondarev, Egor et al.
Eleventh International Conference on Machine Vision, ICMV 2018. ed. / Antanas Verikas; Dmitry P. Nikolaev; Petia Radeva; Jianhong Zhou. Bellingham: SPIE, 2019. 110411C (Proceedings of SPIE; Vol. 11041).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Towards accurate camera geopositioning by image matching

AU - Imbriaco, Raffaele

AU - Sebastian, Clint

AU - Bondarev, Egor

AU - De With, Peter H.N.

N1 - Conference code: 11

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In this work, we present a camera geopositioning system based on matching a query image against a database with panoramic images. For matching, our system uses memory vectors aggregated from global image descriptors based on convolutional features to facilitate fast searching in the database. To speed up searching, a clustering algorithm is used to balance geographical positioning and computation time. We refine the obtained position from the query image using a new outlier removal algorithm. The matching of the query image is obtained with a recall@5 larger than 90% for panorama-to-panorama matching. We cluster available panoramas from geographically adjacent locations into a single compact representation and observe computational gains of approximately 50% at the cost of only a small (approximately 3%) recall loss. Finally, we present a coordinate estimation algorithm that reduces the median geopositioning error by up to 20%.

AB - In this work, we present a camera geopositioning system based on matching a query image against a database with panoramic images. For matching, our system uses memory vectors aggregated from global image descriptors based on convolutional features to facilitate fast searching in the database. To speed up searching, a clustering algorithm is used to balance geographical positioning and computation time. We refine the obtained position from the query image using a new outlier removal algorithm. The matching of the query image is obtained with a recall@5 larger than 90% for panorama-to-panorama matching. We cluster available panoramas from geographically adjacent locations into a single compact representation and observe computational gains of approximately 50% at the cost of only a small (approximately 3%) recall loss. Finally, we present a coordinate estimation algorithm that reduces the median geopositioning error by up to 20%.

KW - convolutional neural networks

KW - geopositioning

KW - Image matching

KW - panoramic images

KW - visual place recognition

UR - http://www.scopus.com/inward/record.url?scp=85063512289&partnerID=8YFLogxK

U2 - 10.1117/12.2522999

DO - 10.1117/12.2522999

M3 - Conference contribution

AN - SCOPUS:85063512289

T3 - Proceedings of SPIE

BT - Eleventh International Conference on Machine Vision, ICMV 2018

A2 - Verikas, Antanas

A2 - Nikolaev, Dmitry P.

A2 - Radeva, Petia

A2 - Zhou, Jianhong

PB - SPIE

CY - Bellingham

T2 - 11th International Conference on Machine Vision, ICMV 2018

Y2 - 1 November 2018 through 3 November 2018

ER -

Towards accurate camera geopositioning by image matching

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this