Fusion of deep learning autoencoders with random forest for wetland classification using Sentinel-2A data: A case study on Sirpur wetland

Notice

This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.

Year : 2026 | Volume : 17 | 01 | Page :
    By

    Reshu Agarwal,

  1. Associate Professor, Department of Computer Science and Engineering Indore Institute of Science and Technology, Rau-Pithampur Road, Opposite IIM, Madhya Pradesh, India

Abstract

Present study analyses the performance of deep leaning algorithm-autoencoder to reduce data dimension as compared to conventional models. Classification accuracies of Sirpur wetland using Sentinel 2A dataset with different inputs have also been studied. These inputs sets comprise the reconstructed data through compression of original 13 bands into 4 bands using decoder algorithm, first four Principal Components, all spectral bands, and spectral indices. Random Forest classifier (RF) is used to classify Sentinel data. Findings revealed that reconstructed input data through decoder has R 2 of 0.99 indicating a very good compression of the original 12 bands data into 4 nodes while Principal Component Analysis (PCA) had 94% of total variability in first four PCs. Classification results showed highest accuracy of 96.25% when nodes were used as input followed by the accuracy of 95% in case of PCs as input. Bands alone could classify with accuracy of 92.5% followed by spectral indices (90.0%). Finally, autoencoder can be a good choice to reduce the number of bands with 2 maximum information especially for complex land surface features and to improve the accuracy. Notably this work is an inaugural step for Sirpur, and results are stepping stone to explore this wonderful ecosystem further.

Keywords: Sirpur, PCA, autoencoder, R software, sentinel-2A

How to cite this article:
Reshu Agarwal. Fusion of deep learning autoencoders with random forest for wetland classification using Sentinel-2A data: A case study on Sirpur wetland. Journal of Remote Sensing & GIS. 2026; 17(01):-.
How to cite this URL:
Reshu Agarwal. Fusion of deep learning autoencoders with random forest for wetland classification using Sentinel-2A data: A case study on Sirpur wetland. Journal of Remote Sensing & GIS. 2026; 17(01):-. Available from: https://journals.stmjournals.com/jorsg/article=2026/view=238892


References

  1. Yuan, S. Y., Liang, X., Lin, T., Chen, S., Liu, R., Wang, J., Zhang, H., & Gong, P. (2025). A comprehensive review of remote sensing in wetlands classification and mapping. arXiv:2504.10842. https://doi.org/10.48550/arXiv.2504.10842
  2. Garg, J. K. (2015). Wetland assessment, monitoring and management in India using geospatial techniques. Journal of Environmental Management, 148, 112-123.
  3. Vaddi, R., Kumar, P. B. L. N., Manoharan, P., Agilandeeswari, L., &  Sangeetha, V. (2024). Strategies for dimensionality reduction in hyperspectral remote sensing: a comprehensive review. The Egyptian Journal of Remote Sensing and Space Science. 27, 82-92.
  4. Kumar, A. &  Garg, R.D., (2023). Assessment and comparison of dimensionality reduction methods on land cover classes and classifiers for PRISMA hyperspectral data. 8th International Conference on Computing in Engineering and Technology (ICCET 2023, 261- 265, doi: 10.1049/icp.2023.1500.
  5. AlSaeed, H., Hewahi, V., & Ksantini, R. (2022). Dimension Reduction Techniques for Image Classification. 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT), Bahrain 358-365, doi: 10.1109/3ICT56508.2022.9990707.
  6. Dronova, I., Gong, P., & Wang, L. (2015). Object-based analysis and change detection of major wetland cover types and their classification uncertainty during low water period at Poyang Lake, China. Remote Sensing of Environment, 115(12), 3220-3236.
  7. Mirmazloumi, S. M., Moghimi, A., Ranjgar, B., Mohseni, F., Ghorbanian, A., Ahmadi, S. A., Amani, M., & Brisco, B. (2021). Status and trends of wetlands studies in Canada using remote sensing technology with a focus on wetland classification: a bibliographic analysis. Remote Sensing, 13(20), 4025.
  8. Aslam, R. W., Naz, I., Shu, H., Yan, J., Quddos, A, Tariq, A., Davis, B., Saif, A. M., & Soufan, W. (2024). Multitemporal image analysis of wetland dynamics using machine learning algorithms. Journal of Environment Management, 371, 123123. 15
  9. Agarwal, R., & Garg, J. K. (2008). Knowledge based classifier of wetlands from coarse resolution satellite data. International Journal of Geoinformatics, 4(3), 17-23.
  10. Agarwal, R., (2017). Model building to investigate the role of spatial location in classifying satellite image using SVM, CART and mBACT: a case study. Journal of Indian Society of Remote Sensing, 45(4), 569–578.
  11. Judah, A., &Hu, B. (2022). An advanced data fusion method to improve wetland classification using multisource remotely sensed data. Sensors, 22, 8942.
  12. Salas, E. A. L., Kumaran, S. S., Bennett, R., Willis, L. P., & Mitchell, K. (2023). Machine learning-based classification of small-sized wetlands using sentinel-2 images. AIIM Geosciences, 10(1), 62-79.
  13. Mohseni, F., Amani, M., Mohammadpour, P., Kakooei, M., Jin, S., & Moghimi, A. (2023). Wetland mapping in great lakes using sentinel-1/2 time-series imagery and DEM data in google earth engine. Remote Sensing, 15(14), 3495. https://doi.org/10.3390/rs15143495.
  14. Jamali, A., Mahdianpari, M., Brisco, B., Granger, J., Mohammadimanesh, F., & Salehi, B. (2021). Deep forest classifier for wetland mapping using the combination of sentinel-1 and sentinel-2 data. GIScience & Remote Sensing, 58 (7), 1072-1089.
  15. Wold, S., Esbensen, K., &  Geladi, P. (1987). Principal component analysis. Chemometrics and Intelligent Laboratory Systems, 2(1), 37-52.
  16. Centeno, J. A., S., Kern, J., Mitishita, E. A., & Palma, M. E. J. (2020). PCA band selection method for hyperspectral sensors onboards an UAV. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Chile IV-3/W2-2020.
  17. R Core Team. (2023). R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria. https://www.R-project.org/. 03 March 2025.
  18. Berahmand, K., Daneshfar, F., Salehi, E. S., Li, Y., & Xu, Y. (2024). Autoencoders and their applications in machine learning: a survey. Artificial Intelligence Review, 28-57, https://doi.org/10.1007/s10462-023-10662-6.
  19. Akar, O., & Gungor, O. (2012). Classification of multispectral images using random forest algorithm. Journal of Geodesy and Geoinformation Science, 1(2), 105-112. 20. Liaw, A., & Wiener, M. (2002). Classification and regression by randomForest. R News, 2(3), 18-22.
Ahead of Print Subscription Original Research
Volume 17
01
Received 22/01/2026
Accepted 02/02/2026
Published 06/02/2026
Publication Time 15 Days
61

Login


My IP

PlumX Metrics