Bedoi: Benchmarks For Determining Overlapping Images With Photogrammetric Information

Publikation: Beitrag in FachzeitschriftKonferenzaufsatz in FachzeitschriftForschungPeer-Review

Autoren

  • H. Zhan
  • Y. F. Yu
  • Q. B. Hou
  • R. Xia
  • Y. Feng
  • Z. Q. Zhan
  • R. Hänsch
  • C. Heipke
  • Michael Gruber
  • Y.W. Xu
  • Xin Wang
  • M.L. Li

Externe Organisationen

  • Wuhan University
  • Technische Universität München (TUM)
  • Deutsches Zentrum für Luft- und Raumfahrt e.V. (DLR)
  • Vexcel Imaging GmbH
  • Nanjing University of Aeronautics and Astronautics
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Seiten (von - bis)1685-1692
Seitenumfang8
FachzeitschriftInternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives
JahrgangXLVIII-1
AusgabenummerW2-2023
PublikationsstatusVeröffentlicht - 14 Dez. 2023
VeranstaltungISPRS Geospatial Week 2023 - Kairo, Ägypten
Dauer: 2 Sept. 20237 Sept. 2023

Abstract

For conventional SfM pipeline, image matching is enduring limitation when considering the time efficiency. In the last few years, to speed up image matching procedure, many image retrieval works were proposed to fast find overlapping image pairs, e.g., bag-of-word that clusters hand-crafted local features in a hierarchical way for efficient similar image retrieval, or learning-based global features (such as, VGG or ResNet) are used to represent image in a global compact manner. However, there are rarely benchmarks with referenced overlapping information to: first, evaluate the retrieval performance; second, fine tune deep-learning models along the direction that is more capable to deal with overlapping image pairs. In this work, based on traditional photogrammetric procedures, relevant photogrammetric information is obtained including image orientation parameters, 3D mesh model and etc., we then generate a benchmark for determining Overlapping Images - BeDOI, in which referenced pairwise overlapping relationships are estimated via rigorous photogrammetric geometry. To extend the generality, in total, BeDOI contains 13667 images which are basically UAV and close-range images of various scene categories, e.g., urban cities, campus, village, historical relics, green land, buildings and etc.

ASJC Scopus Sachgebiete

Zitieren

Bedoi: Benchmarks For Determining Overlapping Images With Photogrammetric Information . / Zhan, H.; Yu, Y. F.; Hou, Q. B. et al.
in: International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, Jahrgang XLVIII-1, Nr. W2-2023, 14.12.2023, S. 1685-1692.

Publikation: Beitrag in FachzeitschriftKonferenzaufsatz in FachzeitschriftForschungPeer-Review

Zhan, H, Yu, YF, Hou, QB, Xia, R, Feng, Y, Zhan, ZQ, Hänsch, R, Heipke, C, Gruber, M, Xu, YW, Wang, X & Li, ML 2023, 'Bedoi: Benchmarks For Determining Overlapping Images With Photogrammetric Information ', International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, Jg. XLVIII-1, Nr. W2-2023, S. 1685-1692. https://doi.org/10.5194/isprs-archives-XLVIII-1-W2-2023-1685-2023
Zhan, H., Yu, Y. F., Hou, Q. B., Xia, R., Feng, Y., Zhan, Z. Q., Hänsch, R., Heipke, C., Gruber, M., Xu, Y. W., Wang, X., & Li, M. L. (2023). Bedoi: Benchmarks For Determining Overlapping Images With Photogrammetric Information . International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, XLVIII-1(W2-2023), 1685-1692. https://doi.org/10.5194/isprs-archives-XLVIII-1-W2-2023-1685-2023
Zhan H, Yu YF, Hou QB, Xia R, Feng Y, Zhan ZQ et al. Bedoi: Benchmarks For Determining Overlapping Images With Photogrammetric Information . International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives. 2023 Dez 14;XLVIII-1(W2-2023):1685-1692. doi: 10.5194/isprs-archives-XLVIII-1-W2-2023-1685-2023
Zhan, H. ; Yu, Y. F. ; Hou, Q. B. et al. / Bedoi : Benchmarks For Determining Overlapping Images With Photogrammetric Information . in: International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives. 2023 ; Jahrgang XLVIII-1, Nr. W2-2023. S. 1685-1692.
Download
@article{6f38e3ad0b6f4c3ca8dd3a8436a00e37,
title = "Bedoi: Benchmarks For Determining Overlapping Images With Photogrammetric Information ",
abstract = "For conventional SfM pipeline, image matching is enduring limitation when considering the time efficiency. In the last few years, to speed up image matching procedure, many image retrieval works were proposed to fast find overlapping image pairs, e.g., bag-of-word that clusters hand-crafted local features in a hierarchical way for efficient similar image retrieval, or learning-based global features (such as, VGG or ResNet) are used to represent image in a global compact manner. However, there are rarely benchmarks with referenced overlapping information to: first, evaluate the retrieval performance; second, fine tune deep-learning models along the direction that is more capable to deal with overlapping image pairs. In this work, based on traditional photogrammetric procedures, relevant photogrammetric information is obtained including image orientation parameters, 3D mesh model and etc., we then generate a benchmark for determining Overlapping Images - BeDOI, in which referenced pairwise overlapping relationships are estimated via rigorous photogrammetric geometry. To extend the generality, in total, BeDOI contains 13667 images which are basically UAV and close-range images of various scene categories, e.g., urban cities, campus, village, historical relics, green land, buildings and etc.",
keywords = "Image Matching, Image Retrieval, Overlapping image pairs, Photogrammetric Information, Structure from Motion (SfM)",
author = "H. Zhan and Yu, {Y. F.} and Hou, {Q. B.} and R. Xia and Y. Feng and Zhan, {Z. Q.} and R. H{\"a}nsch and C. Heipke and Michael Gruber and Y.W. Xu and Xin Wang and M.L. Li",
note = "Funding Information: This work was supported by ISPRS Scientific Initiatives 2023, Natural Science Foundation of Hubei Province, China (No. 2022CFB727) and National Natural Science Foundation of China (No. 61871295). ; ISPRS Geospatial Week 2023 ; Conference date: 02-09-2023 Through 07-09-2023",
year = "2023",
month = dec,
day = "14",
doi = "10.5194/isprs-archives-XLVIII-1-W2-2023-1685-2023",
language = "English",
volume = "XLVIII-1",
pages = "1685--1692",
number = "W2-2023",

}

Download

TY - JOUR

T1 - Bedoi

T2 - ISPRS Geospatial Week 2023

AU - Zhan, H.

AU - Yu, Y. F.

AU - Hou, Q. B.

AU - Xia, R.

AU - Feng, Y.

AU - Zhan, Z. Q.

AU - Hänsch, R.

AU - Heipke, C.

AU - Gruber, Michael

AU - Xu, Y.W.

AU - Wang, Xin

AU - Li, M.L.

N1 - Funding Information: This work was supported by ISPRS Scientific Initiatives 2023, Natural Science Foundation of Hubei Province, China (No. 2022CFB727) and National Natural Science Foundation of China (No. 61871295).

PY - 2023/12/14

Y1 - 2023/12/14

N2 - For conventional SfM pipeline, image matching is enduring limitation when considering the time efficiency. In the last few years, to speed up image matching procedure, many image retrieval works were proposed to fast find overlapping image pairs, e.g., bag-of-word that clusters hand-crafted local features in a hierarchical way for efficient similar image retrieval, or learning-based global features (such as, VGG or ResNet) are used to represent image in a global compact manner. However, there are rarely benchmarks with referenced overlapping information to: first, evaluate the retrieval performance; second, fine tune deep-learning models along the direction that is more capable to deal with overlapping image pairs. In this work, based on traditional photogrammetric procedures, relevant photogrammetric information is obtained including image orientation parameters, 3D mesh model and etc., we then generate a benchmark for determining Overlapping Images - BeDOI, in which referenced pairwise overlapping relationships are estimated via rigorous photogrammetric geometry. To extend the generality, in total, BeDOI contains 13667 images which are basically UAV and close-range images of various scene categories, e.g., urban cities, campus, village, historical relics, green land, buildings and etc.

AB - For conventional SfM pipeline, image matching is enduring limitation when considering the time efficiency. In the last few years, to speed up image matching procedure, many image retrieval works were proposed to fast find overlapping image pairs, e.g., bag-of-word that clusters hand-crafted local features in a hierarchical way for efficient similar image retrieval, or learning-based global features (such as, VGG or ResNet) are used to represent image in a global compact manner. However, there are rarely benchmarks with referenced overlapping information to: first, evaluate the retrieval performance; second, fine tune deep-learning models along the direction that is more capable to deal with overlapping image pairs. In this work, based on traditional photogrammetric procedures, relevant photogrammetric information is obtained including image orientation parameters, 3D mesh model and etc., we then generate a benchmark for determining Overlapping Images - BeDOI, in which referenced pairwise overlapping relationships are estimated via rigorous photogrammetric geometry. To extend the generality, in total, BeDOI contains 13667 images which are basically UAV and close-range images of various scene categories, e.g., urban cities, campus, village, historical relics, green land, buildings and etc.

KW - Image Matching

KW - Image Retrieval

KW - Overlapping image pairs

KW - Photogrammetric Information

KW - Structure from Motion (SfM)

UR - http://www.scopus.com/inward/record.url?scp=85183290467&partnerID=8YFLogxK

U2 - 10.5194/isprs-archives-XLVIII-1-W2-2023-1685-2023

DO - 10.5194/isprs-archives-XLVIII-1-W2-2023-1685-2023

M3 - Conference article

AN - SCOPUS:85183290467

VL - XLVIII-1

SP - 1685

EP - 1692

JO - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives

JF - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives

SN - 1682-1750

IS - W2-2023

Y2 - 2 September 2023 through 7 September 2023

ER -