A Comparative Performance Analysis of Popular Deep Learning Models and Segment Anything Model (SAM) for River Water Segmentation in Close-Range Remote Sensing Imagery: English

Publikation: Beitrag in FachzeitschriftArtikelForschungPeer-Review

Autoren

Externe Organisationen

  • University of the Witwatersrand
  • University of Agder
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Seiten (von - bis)52067-52085
Seitenumfang19
FachzeitschriftIEEE ACCESS
Jahrgang12
PublikationsstatusVeröffentlicht - 5 Apr. 2024

Abstract

Accurate segmentation of river water in close-range Remote Sensing (RS) images is vital for efficient environmental monitoring and management. However, this task poses significant difficulties due to the dynamic nature of water, which exhibits varying colors and textures reflecting the sky and surrounding structures along the riverbanks. This study addresses these complexities by evaluating and comparing several well-known deep-learning (DL) techniques on four river scene datasets. To achieve this, we fine-tuned the recently introduced "Segment Anything Model" (SAM) along with popular DL segmentation models such as U-Net, DeepLabV3+, LinkNet, PSPNet, and PAN, all using ResNet50 pre-trained on ImageNet as a backbone. Experimental results highlight the diverse performances of these models in river water segmentation. Notably, fine-tuned SAM demonstrates superior performance, followed by U-Net(ResNet50), despite their higher computational costs. In contrast, PSPNet(ResNet50), while less effective, proves to be the most efficient in terms of execution time. In addition to these findings, we introduce a novel river water segmentation dataset, LuFI-RiverSnap.<italic>v</italic>1 (Dataset link), characterized by a more diverse range of scenes and accurate masks compared to existing datasets. To facilitate reproducible research in remote sensing and computer vision, we release the implementations of the fine-tuned SAM model (Code link). The findings from this research, coupled with the presented dataset and the accuracy achieved by fine-tuned SAM segmentation, can support tracking river changes, understanding river water level trends, and exploring river ecosystem dynamics. These can also provide valuable insights for practitioners and researchers seeking models tailored to specific image characteristics with practical means in disaster risk reduction, such as rapid assessments of inundations during floods or automatic extractions of gauge data in watersheds.

Zitieren

A Comparative Performance Analysis of Popular Deep Learning Models and Segment Anything Model (SAM) for River Water Segmentation in Close-Range Remote Sensing Imagery: English. / Moghimi, Armin; Welzel, Mario; Celik, Turgay et al.
in: IEEE ACCESS, Jahrgang 12, 05.04.2024, S. 52067-52085.

Publikation: Beitrag in FachzeitschriftArtikelForschungPeer-Review

Download
@article{b4dbc46b28ee407b854dad3e25da0fe0,
title = "A Comparative Performance Analysis of Popular Deep Learning Models and Segment Anything Model (SAM) for River Water Segmentation in Close-Range Remote Sensing Imagery: English",
abstract = "Accurate segmentation of river water in close-range Remote Sensing (RS) images is vital for efficient environmental monitoring and management. However, this task poses significant difficulties due to the dynamic nature of water, which exhibits varying colors and textures reflecting the sky and surrounding structures along the riverbanks. This study addresses these complexities by evaluating and comparing several well-known deep-learning (DL) techniques on four river scene datasets. To achieve this, we fine-tuned the recently introduced {"}Segment Anything Model{"} (SAM) along with popular DL segmentation models such as U-Net, DeepLabV3+, LinkNet, PSPNet, and PAN, all using ResNet50 pre-trained on ImageNet as a backbone. Experimental results highlight the diverse performances of these models in river water segmentation. Notably, fine-tuned SAM demonstrates superior performance, followed by U-Net(ResNet50), despite their higher computational costs. In contrast, PSPNet(ResNet50), while less effective, proves to be the most efficient in terms of execution time. In addition to these findings, we introduce a novel river water segmentation dataset, LuFI-RiverSnap.v1 (Dataset link), characterized by a more diverse range of scenes and accurate masks compared to existing datasets. To facilitate reproducible research in remote sensing and computer vision, we release the implementations of the fine-tuned SAM model (Code link). The findings from this research, coupled with the presented dataset and the accuracy achieved by fine-tuned SAM segmentation, can support tracking river changes, understanding river water level trends, and exploring river ecosystem dynamics. These can also provide valuable insights for practitioners and researchers seeking models tailored to specific image characteristics with practical means in disaster risk reduction, such as rapid assessments of inundations during floods or automatic extractions of gauge data in watersheds.",
keywords = "Biological system modeling, Cameras, Deep learning, DeepLabV3+, Feature extraction, Image segmentation, LinkNet, PAN, PSPNet, Residual neural networks, river water segmentation, Rivers, RiverSnap, Segment Anything Model (SAM), Task analysis, U-Net, segment anything model (SAM)",
author = "Armin Moghimi and Mario Welzel and Turgay Celik and Torsten Schlurmann",
year = "2024",
month = apr,
day = "5",
doi = "10.1109/ACCESS.2024.3385425",
language = "English",
volume = "12",
pages = "52067--52085",
journal = "IEEE ACCESS",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Download

TY - JOUR

T1 - A Comparative Performance Analysis of Popular Deep Learning Models and Segment Anything Model (SAM) for River Water Segmentation in Close-Range Remote Sensing Imagery

T2 - English

AU - Moghimi, Armin

AU - Welzel, Mario

AU - Celik, Turgay

AU - Schlurmann, Torsten

PY - 2024/4/5

Y1 - 2024/4/5

N2 - Accurate segmentation of river water in close-range Remote Sensing (RS) images is vital for efficient environmental monitoring and management. However, this task poses significant difficulties due to the dynamic nature of water, which exhibits varying colors and textures reflecting the sky and surrounding structures along the riverbanks. This study addresses these complexities by evaluating and comparing several well-known deep-learning (DL) techniques on four river scene datasets. To achieve this, we fine-tuned the recently introduced "Segment Anything Model" (SAM) along with popular DL segmentation models such as U-Net, DeepLabV3+, LinkNet, PSPNet, and PAN, all using ResNet50 pre-trained on ImageNet as a backbone. Experimental results highlight the diverse performances of these models in river water segmentation. Notably, fine-tuned SAM demonstrates superior performance, followed by U-Net(ResNet50), despite their higher computational costs. In contrast, PSPNet(ResNet50), while less effective, proves to be the most efficient in terms of execution time. In addition to these findings, we introduce a novel river water segmentation dataset, LuFI-RiverSnap.v1 (Dataset link), characterized by a more diverse range of scenes and accurate masks compared to existing datasets. To facilitate reproducible research in remote sensing and computer vision, we release the implementations of the fine-tuned SAM model (Code link). The findings from this research, coupled with the presented dataset and the accuracy achieved by fine-tuned SAM segmentation, can support tracking river changes, understanding river water level trends, and exploring river ecosystem dynamics. These can also provide valuable insights for practitioners and researchers seeking models tailored to specific image characteristics with practical means in disaster risk reduction, such as rapid assessments of inundations during floods or automatic extractions of gauge data in watersheds.

AB - Accurate segmentation of river water in close-range Remote Sensing (RS) images is vital for efficient environmental monitoring and management. However, this task poses significant difficulties due to the dynamic nature of water, which exhibits varying colors and textures reflecting the sky and surrounding structures along the riverbanks. This study addresses these complexities by evaluating and comparing several well-known deep-learning (DL) techniques on four river scene datasets. To achieve this, we fine-tuned the recently introduced "Segment Anything Model" (SAM) along with popular DL segmentation models such as U-Net, DeepLabV3+, LinkNet, PSPNet, and PAN, all using ResNet50 pre-trained on ImageNet as a backbone. Experimental results highlight the diverse performances of these models in river water segmentation. Notably, fine-tuned SAM demonstrates superior performance, followed by U-Net(ResNet50), despite their higher computational costs. In contrast, PSPNet(ResNet50), while less effective, proves to be the most efficient in terms of execution time. In addition to these findings, we introduce a novel river water segmentation dataset, LuFI-RiverSnap.v1 (Dataset link), characterized by a more diverse range of scenes and accurate masks compared to existing datasets. To facilitate reproducible research in remote sensing and computer vision, we release the implementations of the fine-tuned SAM model (Code link). The findings from this research, coupled with the presented dataset and the accuracy achieved by fine-tuned SAM segmentation, can support tracking river changes, understanding river water level trends, and exploring river ecosystem dynamics. These can also provide valuable insights for practitioners and researchers seeking models tailored to specific image characteristics with practical means in disaster risk reduction, such as rapid assessments of inundations during floods or automatic extractions of gauge data in watersheds.

KW - Biological system modeling

KW - Cameras

KW - Deep learning

KW - DeepLabV3+

KW - Feature extraction

KW - Image segmentation

KW - LinkNet

KW - PAN

KW - PSPNet

KW - Residual neural networks

KW - river water segmentation

KW - Rivers

KW - RiverSnap

KW - Segment Anything Model (SAM)

KW - Task analysis

KW - U-Net

KW - segment anything model (SAM)

UR - http://www.scopus.com/inward/record.url?scp=85189815510&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2024.3385425

DO - 10.1109/ACCESS.2024.3385425

M3 - Article

AN - SCOPUS:85189815510

VL - 12

SP - 52067

EP - 52085

JO - IEEE ACCESS

JF - IEEE ACCESS

SN - 2169-3536

ER -