A Method for Extracting Lake Water Using ViTenc-UNet: Taking Typical Lakes on the Qinghai-Tibet Plateau as Examples DOI Creative Commons

Xili Zhao,

Hong Wang, Li Liu

et al.

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(16), P. 4047 - 4047

Published: Aug. 16, 2023

As the lakes located in Qinghai-Tibet Plateau are important carriers of water resources Asia, dynamic changes to these intuitively reflect climate and resource variations Plateau. To address insufficient performance Convolutional Neural Network (CNN) learning spatial relationship between long-distance continuous pixels, this study proposes a recognition model for on based U-Net ViTenc-UNet. This method uses Vision Transformer (ViT) replace layer encoder model, which can more accurately identify extract lake bodies. A Block Attention Module (CBAM) mechanism was added decoder enabling information spectral characteristics bodies be completely preserved. The experimental results show that ViTenc-UNet complete task efficiently, Overall Accuracy, Intersection over Union, Recall, Precision, F1 score classification reached 99.04%, 98.68%, 99.08%, 98.59%, 98.75%, were, respectively, 4.16%, 6.20% 5.34%, 4.80%, 5.34% higher than original model. Compared FCN, DeepLabv3+, TransUNet, Swin-Unet models also have different degrees advantages. innovatively introduces ViT CBAM into extraction Plateau, showing excellent has certain advantages will provide an scientific reference accurate real-time monitoring

Language: Английский

CPVF: vectorization of agricultural cultivation field parcels via a boundary–parcel multi-task learning network in ultra-high-resolution remote sensing images DOI
Xiuyu Liu,

Jinshui Zhang,

Yaming Duan

et al.

ISPRS Journal of Photogrammetry and Remote Sensing, Journal Year: 2025, Volume and Issue: 226, P. 267 - 299

Published: May 26, 2025

Language: Английский

Citations

0
Unsupervised Deep Learning for Flood Segmentation in UAV Imagery DOI
Georgios Simantiris, Costas Panagiotakis

Lecture notes in computer science, Journal Year: 2025, Volume and Issue: unknown, P. 408 - 423

Published: Jan. 1, 2025

Language: Английский

Citations

0
Improvement and application of UNet network for avoiding the effect of urban dense high-rise buildings and other feature shadows on water body extraction DOI
Yiheng Xie, Renxi Chen,

Mingge Yu

et al.

International Journal of Remote Sensing, Journal Year: 2023, Volume and Issue: 44(12), P. 3861 - 3891

Published: June 18, 2023

ABSTRACTFinding a means to extract water body information efficiently and accurately from high-resolution remote sensing images has been an important research direction in the field of extraction recent years. However, shadows buildings other obstacles interfere with accuracy extraction. To address this problem, paper proposes neural network method incorporating attention mechanism for This is based on U-Net convolutional adds squeeze-and-excitation module SENet, mechanism, downsampling process network. The weights feature maps so that focuses more features thus reduces shadow features, improving image segmentation. dropout batch normalization layers are also added improve generalization ability stability model. In paper, SE-CU-Net model presented overcome shadowing effect features. Using GF-2 Jiangsu province as data source, recognition results compared Dense-Net, Res-Net, Seg-Net, U-net, SVM, RF. Through comparison experiments, can not only better influence but it stronger effect. average ASCR, Precision, mIoU, OA, F1-Score kappa coefficients three tested areas reached 98.27%, 97.17%, 89.33%, 98.2%, 89.3% 0.883, respectively, significantly higher than six classical methods, verifying effectiveness overcoming research.KEYWORDS: extractiondeep learningshadows buildingsU-Netattention Disclosure statementNo potential conflict interest was reported by author(s).Additional informationFundingThis work Funded Key Laboratory Land Satellite Remote Sensing Application, Ministry Natural Resources People's Republic China(Grant No. KLSMNR-G202212)

Language: Английский

Citations

8
A novel convolutional neural network model with hybrid attentional atrous convolution module for detecting the areas affected by the flood DOI
Abdullah ŞENER, Gürkan Doğan, Burhan Ergen

et al.

Earth Science Informatics, Journal Year: 2023, Volume and Issue: 17(1), P. 193 - 209

Published: Nov. 24, 2023

Language: Английский

Citations

8
A Method for Extracting Lake Water Using ViTenc-UNet: Taking Typical Lakes on the Qinghai-Tibet Plateau as Examples DOI Creative Commons

Xili Zhao,

Hong Wang, Li Liu

et al.

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(16), P. 4047 - 4047

Published: Aug. 16, 2023

As the lakes located in Qinghai-Tibet Plateau are important carriers of water resources Asia, dynamic changes to these intuitively reflect climate and resource variations Plateau. To address insufficient performance Convolutional Neural Network (CNN) learning spatial relationship between long-distance continuous pixels, this study proposes a recognition model for on based U-Net ViTenc-UNet. This method uses Vision Transformer (ViT) replace layer encoder model, which can more accurately identify extract lake bodies. A Block Attention Module (CBAM) mechanism was added decoder enabling information spectral characteristics bodies be completely preserved. The experimental results show that ViTenc-UNet complete task efficiently, Overall Accuracy, Intersection over Union, Recall, Precision, F1 score classification reached 99.04%, 98.68%, 99.08%, 98.59%, 98.75%, were, respectively, 4.16%, 6.20% 5.34%, 4.80%, 5.34% higher than original model. Compared FCN, DeepLabv3+, TransUNet, Swin-Unet models also have different degrees advantages. innovatively introduces ViT CBAM into extraction Plateau, showing excellent has certain advantages will provide an scientific reference accurate real-time monitoring

Language: Английский

Citations

7