Unsupervised vehicle extraction of bounding boxes in UAV images

Junho Yeom; Youkyung Han

doi:10.1117/12.2680067

Unsupervised vehicle extraction of bounding boxes in UAV images

Junho Yeom
, Youkyung Han

Dept. of Civil Engineering

Gyeongsang National University

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

Abstract

Various studies have been conducted to detect objects in urban areas by applying machine learning algorithms to UAV high-resolution images. However, most vehicle detection studies have limitations in that vehicle detection is performed as a bounding box instead of instance segmentation. Since instance segmentation requires labor-intensive labeling work of each object to train individual objects, research on how to perform unsupervised automatic instance segmentation is needed. Therefore, this study proposed unsupervised SVM classification of the vehicle bounding boxes in UAV images for instance segmentation. As a result of the extraction, it was confirmed that the vehicle could be detected with an accuracy of 89%. It was also confirmed that the vehicle could be detected even if the spectral characteristics within the vehicle were significantly different.

Original language	English
Title of host publication	Remote Sensing Technologies and Applications in Urban Environments VIII
Editors	Thilo Erbertseder, Nektarios Chrysoulakis, Ying Zhang
Publisher	SPIE
ISBN (Electronic)	9781510666993
DOIs	https://doi.org/10.1117/12.2680067
State	Published - 2023
Event	Remote Sensing Technologies and Applications in Urban Environments VIII 2023 - Amsterdam, Netherlands Duration: 3 Sep 2023 → 4 Sep 2023

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	12735
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	Remote Sensing Technologies and Applications in Urban Environments VIII 2023
Country/Territory	Netherlands
City	Amsterdam
Period	3/09/23 → 4/09/23

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 11 Sustainable Cities and Communities

Keywords

UAV
Unsupervised SVM
Vehicle extraction

Access to Document

10.1117/12.2680067

Cite this

@inproceedings{ebbf60c2b853471b954a8da53cfababd,

title = "Unsupervised vehicle extraction of bounding boxes in UAV images",

abstract = "Various studies have been conducted to detect objects in urban areas by applying machine learning algorithms to UAV high-resolution images. However, most vehicle detection studies have limitations in that vehicle detection is performed as a bounding box instead of instance segmentation. Since instance segmentation requires labor-intensive labeling work of each object to train individual objects, research on how to perform unsupervised automatic instance segmentation is needed. Therefore, this study proposed unsupervised SVM classification of the vehicle bounding boxes in UAV images for instance segmentation. As a result of the extraction, it was confirmed that the vehicle could be detected with an accuracy of 89\%. It was also confirmed that the vehicle could be detected even if the spectral characteristics within the vehicle were significantly different.",

keywords = "UAV, Unsupervised SVM, Vehicle extraction",

author = "Junho Yeom and Youkyung Han",

note = "Publisher Copyright: {\textcopyright} 2023 SPIE.; Remote Sensing Technologies and Applications in Urban Environments VIII 2023 ; Conference date: 03-09-2023 Through 04-09-2023",

year = "2023",

doi = "10.1117/12.2680067",

language = "English",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

editor = "Thilo Erbertseder and Nektarios Chrysoulakis and Ying Zhang",

booktitle = "Remote Sensing Technologies and Applications in Urban Environments VIII",

}

Yeom, J & Han, Y 2023, Unsupervised vehicle extraction of bounding boxes in UAV images. in T Erbertseder, N Chrysoulakis & Y Zhang (eds), Remote Sensing Technologies and Applications in Urban Environments VIII., 127350N, Proceedings of SPIE - The International Society for Optical Engineering, vol. 12735, SPIE, Remote Sensing Technologies and Applications in Urban Environments VIII 2023, Amsterdam, Netherlands, 3/09/23. https://doi.org/10.1117/12.2680067

Unsupervised vehicle extraction of bounding boxes in UAV images. / Yeom, Junho; Han, Youkyung.
Remote Sensing Technologies and Applications in Urban Environments VIII. ed. / Thilo Erbertseder; Nektarios Chrysoulakis; Ying Zhang. SPIE, 2023. 127350N (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 12735).

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

TY - GEN

T1 - Unsupervised vehicle extraction of bounding boxes in UAV images

AU - Yeom, Junho

AU - Han, Youkyung

PY - 2023

Y1 - 2023

N2 - Various studies have been conducted to detect objects in urban areas by applying machine learning algorithms to UAV high-resolution images. However, most vehicle detection studies have limitations in that vehicle detection is performed as a bounding box instead of instance segmentation. Since instance segmentation requires labor-intensive labeling work of each object to train individual objects, research on how to perform unsupervised automatic instance segmentation is needed. Therefore, this study proposed unsupervised SVM classification of the vehicle bounding boxes in UAV images for instance segmentation. As a result of the extraction, it was confirmed that the vehicle could be detected with an accuracy of 89%. It was also confirmed that the vehicle could be detected even if the spectral characteristics within the vehicle were significantly different.

AB - Various studies have been conducted to detect objects in urban areas by applying machine learning algorithms to UAV high-resolution images. However, most vehicle detection studies have limitations in that vehicle detection is performed as a bounding box instead of instance segmentation. Since instance segmentation requires labor-intensive labeling work of each object to train individual objects, research on how to perform unsupervised automatic instance segmentation is needed. Therefore, this study proposed unsupervised SVM classification of the vehicle bounding boxes in UAV images for instance segmentation. As a result of the extraction, it was confirmed that the vehicle could be detected with an accuracy of 89%. It was also confirmed that the vehicle could be detected even if the spectral characteristics within the vehicle were significantly different.

KW - UAV

KW - Unsupervised SVM

KW - Vehicle extraction

UR - https://www.scopus.com/pages/publications/85180007064

U2 - 10.1117/12.2680067

DO - 10.1117/12.2680067

M3 - Conference contribution

AN - SCOPUS:85180007064

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Remote Sensing Technologies and Applications in Urban Environments VIII

A2 - Erbertseder, Thilo

A2 - Chrysoulakis, Nektarios

A2 - Zhang, Ying

PB - SPIE

T2 - Remote Sensing Technologies and Applications in Urban Environments VIII 2023

Y2 - 3 September 2023 through 4 September 2023

ER -

Unsupervised vehicle extraction of bounding boxes in UAV images

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this