Classification of estuaries and coastal wetlands from Planet-NICFI imagery based on convolutional neural networks and transfer training

Remote sensing

UDC:

528.7:528.8

DOI:

10.22389/0016-7126-2024-1008-6-31-42

1 Quyen D.T.

2 Malinnikov V.A.

Year:

2024

№:

1008

Pages:

31-42

Moscow State University of Geodesy and Cartography (MIIGAiK)

1,

2,

Abstract:

The authors consider the importance of monitoring coastal wetland ecosystems, negatively impacted by human activities and climate change. In this context, artificial intelligence neural networks are applied to classify this type of wetland. However, they encounter a task that requires extensive volume of training data to achieve high accuracy results. Within the conducted research, a method of transfer training from neural networks is proposed to overcome the aforementioned problem. The developed model combines multi-temporal Planet-NICFI satellite images for classifying coastal wetlands, especially under tidal conditions. The research results indicate that the model has upgraded its accuracy from 89,2 % to 91,3 % in the wetlands of the Ba Lat estuary. Besides, it has been successfully applied to classify similar lands in the Red River Biosphere Reserve during the period of 2016–2022. This will enable improving the management of this area in the future

Keywords:

Ba Lat estuary, coastal wetlands, deep learning, neural network, Red River delta, transfer learning, U-Net

The work was carried out within the framework of the state assignment No. FSFE-2023-0005. The training and testing of the neural network classification model were conducted within the framework of the state assignment No. FSFE-2022-0003

References:

1. Kuen Din' Tuen, Malinnikov V. A., Fam Chi Kong Razrabotka metodiki opredeleniya tipov pribrezhnykh vodno-bolotnykh ugodii po kosmicheskim izobrazheniyam Planetscope s ispol'zovaniem metoda glubokogo obucheniya (na primere natsional'nogo parka Mui Kamau, V'etnam). Izvestia vuzov. Geodesy and Aerophotosurveying, 2023, Vol. 67, no. 6, pp. 109–122.

2. Dang K. B., Nguyen M. H., Nguyen D. A., et al. (2020) Coastal wetland classification with deep U-Net convolutional networks and sentinel-2 imagery: A case study at the Tien Yen estuary of Vietnam. Remote Sensing, DOI: 10.3390/rs12193270.

3. Guan X., Wang D., Wan L., Zhang J. (2022) Extracting Wetland Type Information with a Deep Convolutional Neural Network. Computational Intelligence and Neuroscience, no. 9, pp. 1–11. DOI: 10.1155/2022/5303872.

4. He K., Zhang X., Ren S., Sun J. (2016) ResNet34: Deep Residual Learning for Image Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 770–778. DOI: 10.1109/CVPR.2016.90.

5. Hopkinson C. S., Wolanski E., Cahoon D. R., Perillo G. M., Brinson M. M. (2019) Coastal wetlands: a synthesis. In book: Coastal Wetlands: An Integrated Ecosystem Approach (Second Edition). pp. 1–75. DOI: 10.1016/B978-0-444-63893-9.00001-0.

6. Kingma D. P., Ba J. (2015) Adam: A Method for Stochastic Optimization. URL: arxiv.org/abs/1412.6980 (accessed: 10.10.2022). DOI: 10.48550/arXiv.1412.6980.

7. Ma Y., Chen S., Ermon S., Lobell D. B. (2024) Transfer learning in environmental remote sensing. Remote Sensing of Environment, no. 301: 113924, DOI: 10.1016/j.rse.2023.113924.

8. Pedersen A., Thang Nguyen Huy, Dung Vu Van, Tri Hoang Trong (1996) The Conservation of Key Coastal Wetland Sites in the Red River Delta. BirdLife International Vietnam Programme, Hanoi, 97 p.

9. Ronneberger O., Fischer P., Brox T. (2015) U-Net: Convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Lecture Notes in Computer Science. DOI: 10.1007/978-3-319-24574-4_28.

10. Slayde Hawkins, Chu Cuong (2010) Roots in the Water: Legal Frameworks for Mangrove PES in Vietnam. Forest Trends and Katoomba Group, Washington, 46 p.

11. Vermote E. F., Tanré D., Deuze J., Herman M., Morcrette J.-J. (1997) Second Simulation of the Satellite Signal in the Solar Spectrum, 6S. IEEE transactions on geoscience and remote sensing, no. 35 (3), pp. 675–686. DOI: 10.1109/36.581987.

Citation:

Quyen D.T.,

Malinnikov V.A.,

(2024) Classification of estuaries and coastal wetlands from Planet-NICFI imagery based on convolutional neural networks and transfer training. Geodesy and cartography = Geodezia i Kartografia, 85(6), pp. 31-42. (In Russian). DOI: 10.22389/0016-7126-2024-1008-6-31-42