A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images

Muhammet Tahir Güneşer; Chihat Şeker; Mohammed Ayad Alkhafaji

doi:10.1007/978-981-97-8160-7_27

A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images

Muhammet Tahir Güneşer, Chihat Şeker, Mohammed Ayad Alkhafaji^*

^*Corresponding author for this work

Department of Electronics and Communication Engineering

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

Abstract

Image segmentation at the pixel level is a time-consuming and difficult task in computer vision and image processing. Aerial (satellite/drone) photo segmentation is considered in this paper. Data from high-resolution remote sensing has enabled new applications such as more detailed per-pixel object classification. U-Net with VGG16 has made segmentation and categorization of images much more efficient and intelligent. U-Net models with pre-trained VGG16 backbones perform best across all tested scenarios. Adding the near-infrared band improves prediction results slightly compared with using RGB image bands alone. The ability to transfer images between sensors, especially between satellites and aerial images, could be improved through train-time enhancement and contrast enhancement. Further improving performance could be achieved by adding noisy training data from free online resources.

Original language	English
Title of host publication	Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024
Editors	Siba Kumar Udgata, Srinivas Sethi, George Ghinea, Sanjay Kumar Kuanar
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	349-362
Number of pages	14
ISBN (Print)	9789819781591
DOIs	https://doi.org/10.1007/978-981-97-8160-7_27
Publication status	Published - 2025
Event	4th International Conference on Machine Learning, Internet of Things and Big Data, ICMIB 2024 - Gunupur, India Duration: 8 Mar 2024 → 10 Mar 2024

Publication series

Name	Lecture Notes in Networks and Systems
Volume	1149
ISSN (Print)	2367-3370
ISSN (Electronic)	2367-3389

Conference

Conference	4th International Conference on Machine Learning, Internet of Things and Big Data, ICMIB 2024
Country/Territory	India
City	Gunupur
Period	8/03/24 → 10/03/24

Bibliographical note

Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

Keywords

Deep learning
Drone images
U-Net

Access to Document

10.1007/978-981-97-8160-7_27

Cite this

Güneşer, M. T., Şeker, C., & Alkhafaji, M. A. (2025). A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images. In S. Kumar Udgata, S. Sethi, G. Ghinea, & S. K. Kuanar (Eds.), Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024 (pp. 349-362). (Lecture Notes in Networks and Systems; Vol. 1149). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-97-8160-7_27

Güneşer, Muhammet Tahir ; Şeker, Chihat ; Alkhafaji, Mohammed Ayad. / A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images. Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024. editor / Siba Kumar Udgata ; Srinivas Sethi ; George Ghinea ; Sanjay Kumar Kuanar. Springer Science and Business Media Deutschland GmbH, 2025. pp. 349-362 (Lecture Notes in Networks and Systems).

@inproceedings{f26c1df3045e4d7893e25ee21e209a69,

title = "A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images",

abstract = "Image segmentation at the pixel level is a time-consuming and difficult task in computer vision and image processing. Aerial (satellite/drone) photo segmentation is considered in this paper. Data from high-resolution remote sensing has enabled new applications such as more detailed per-pixel object classification. U-Net with VGG16 has made segmentation and categorization of images much more efficient and intelligent. U-Net models with pre-trained VGG16 backbones perform best across all tested scenarios. Adding the near-infrared band improves prediction results slightly compared with using RGB image bands alone. The ability to transfer images between sensors, especially between satellites and aerial images, could be improved through train-time enhancement and contrast enhancement. Further improving performance could be achieved by adding noisy training data from free online resources.",

keywords = "Deep learning, Drone images, U-Net",

author = "G{\"u}ne{\c s}er, {Muhammet Tahir} and Chihat {\c S}eker and Alkhafaji, {Mohammed Ayad}",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.; 4th International Conference on Machine Learning, Internet of Things and Big Data, ICMIB 2024 ; Conference date: 08-03-2024 Through 10-03-2024",

year = "2025",

doi = "10.1007/978-981-97-8160-7_27",

language = "English",

isbn = "9789819781591",

series = "Lecture Notes in Networks and Systems",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "349--362",

editor = "{Kumar Udgata}, Siba and Srinivas Sethi and George Ghinea and Kuanar, {Sanjay Kumar}",

booktitle = "Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024",

address = "Germany",

}

Güneşer, MT, Şeker, C & Alkhafaji, MA 2025, A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images. in S Kumar Udgata, S Sethi, G Ghinea & SK Kuanar (eds), Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024. Lecture Notes in Networks and Systems, vol. 1149, Springer Science and Business Media Deutschland GmbH, pp. 349-362, 4th International Conference on Machine Learning, Internet of Things and Big Data, ICMIB 2024, Gunupur, India, 8/03/24. https://doi.org/10.1007/978-981-97-8160-7_27

A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images. / Güneşer, Muhammet Tahir; Şeker, Chihat; Alkhafaji, Mohammed Ayad.
Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024. ed. / Siba Kumar Udgata; Srinivas Sethi; George Ghinea; Sanjay Kumar Kuanar. Springer Science and Business Media Deutschland GmbH, 2025. p. 349-362 (Lecture Notes in Networks and Systems; Vol. 1149).

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

TY - GEN

T1 - A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images

AU - Güneşer, Muhammet Tahir

AU - Şeker, Chihat

AU - Alkhafaji, Mohammed Ayad

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

PY - 2025

Y1 - 2025

N2 - Image segmentation at the pixel level is a time-consuming and difficult task in computer vision and image processing. Aerial (satellite/drone) photo segmentation is considered in this paper. Data from high-resolution remote sensing has enabled new applications such as more detailed per-pixel object classification. U-Net with VGG16 has made segmentation and categorization of images much more efficient and intelligent. U-Net models with pre-trained VGG16 backbones perform best across all tested scenarios. Adding the near-infrared band improves prediction results slightly compared with using RGB image bands alone. The ability to transfer images between sensors, especially between satellites and aerial images, could be improved through train-time enhancement and contrast enhancement. Further improving performance could be achieved by adding noisy training data from free online resources.

AB - Image segmentation at the pixel level is a time-consuming and difficult task in computer vision and image processing. Aerial (satellite/drone) photo segmentation is considered in this paper. Data from high-resolution remote sensing has enabled new applications such as more detailed per-pixel object classification. U-Net with VGG16 has made segmentation and categorization of images much more efficient and intelligent. U-Net models with pre-trained VGG16 backbones perform best across all tested scenarios. Adding the near-infrared band improves prediction results slightly compared with using RGB image bands alone. The ability to transfer images between sensors, especially between satellites and aerial images, could be improved through train-time enhancement and contrast enhancement. Further improving performance could be achieved by adding noisy training data from free online resources.

KW - Deep learning

KW - Drone images

KW - U-Net

UR - http://www.scopus.com/inward/record.url?scp=85218224704&partnerID=8YFLogxK

U2 - 10.1007/978-981-97-8160-7_27

DO - 10.1007/978-981-97-8160-7_27

M3 - Conference contribution

AN - SCOPUS:85218224704

SN - 9789819781591

T3 - Lecture Notes in Networks and Systems

SP - 349

EP - 362

BT - Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024

A2 - Kumar Udgata, Siba

A2 - Sethi, Srinivas

A2 - Ghinea, George

A2 - Kuanar, Sanjay Kumar

PB - Springer Science and Business Media Deutschland GmbH

T2 - 4th International Conference on Machine Learning, Internet of Things and Big Data, ICMIB 2024

Y2 - 8 March 2024 through 10 March 2024

ER -

Güneşer MT, Şeker C, Alkhafaji MA. A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images. In Kumar Udgata S, Sethi S, Ghinea G, Kuanar SK, editors, Intelligent Systems - Proceedings of 4th International Conference on Machine Learning, IoT and Big Data ICMIB 2024. Springer Science and Business Media Deutschland GmbH. 2025. p. 349-362. (Lecture Notes in Networks and Systems). doi: 10.1007/978-981-97-8160-7_27

A Real-Time Aerial Semantic Segmentation System Based on U-Net Deep Learning Using Drone Images

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this