Managing Uncertainty in Land Use Change Detection: A Comparative Analysis of Classical and Modern Machine Learning Approaches

Document Type : Original Article

Authors

Faculty of Civil, Water, and Environmental Engineering, Shahid Beheshti University, Tehran 1983969411, Iran

Abstract

Land use change detection is critical for sustainable environmental management, yet uncertainties from noise, mixed pixels, and spectral similarities challenge its accuracy. This study conducts a comparative analysis of classical machine learning methods—Support Vector Machines, Random Forests, and Maximum Likelihood classifiers—and modern approaches, specifically Convolutional Neural Networks and Bayesian Neural Networks, to evaluate their efficacy in managing uncertainty across urban, agricultural, and aquatic contexts. Utilizing global and Iranian case studies, the research assesses performance metrics, including accuracy, uncertainty management, and computational complexity, through quantitative and qualitative syntheses. Findings reveal that modern methods outperform classical approaches, with Convolutional Neural Networks achieving 90–95% accuracy and Bayesian Neural Networks reaching 91.85% in urban settings, driven by robust feature extraction and probabilistic uncertainty quantification. Classical methods, while less accurate (65–92%), offer computational efficiency, making them viable in resource-constrained regions. The study highlights practical implications for Iran’s urban and agricultural monitoring and global sustainability goals, proposing hybrid approaches and multi-modal data integration to balance accuracy and accessibility. Despite their potential, challenges such as computational intensity, data scarcity, and model interpretability persist, necessitating future research into lightweight algorithms, semi-supervised learning, and explainable artificial intelligence. This analysis advances the field by providing a framework for method selection, enhancing the reliability of land use change detection for environmental policy and resource management.

Keywords

References

Ahmadpour, A., Soleimani, K., Shokri, M., & Ghorbani, J. (2014). Comparison of the efficiency of three common supervised classification methods of satellite data in vegetation cover studies. Civilica. https://civilica.com/doc/1166233
Akhbari, M., Ranjbar, A., & Fatemi, S. M. B. (2006). Investigation of satellite image classification methods. Civilica. https://civilica.com/doc/1389513
Cao, C., Dragićević, S., & Li, S. (2019). Land-use change detection with convolutional neural network methods. Environments, 6(2), 25. https://doi.org/10.3390/environments6020025
Chen, Y., Li, X., & Zhang, S. (2020). Uncertainty analysis in land cover classification using deep learning. Remote Sensing, 12(15), 2345. https://doi.org/10.3390/rs12152345
Foody, G. M. (2010). Assessing the accuracy of land cover change with imperfect ground reference data. Remote Sensing of Environment, 114(10), 2271-2285. https://doi.org/10.1016/j.rse.2010.05.003
Gal, Y., & Ghahramani, Z. (2016). Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. Proceedings of the 33rd International Conference on Machine Learning, 48, 1050-1059.
Huang, C., Davis, L. S., & Townshend, J. R. G. (2002). An assessment of support vector machines for land cover classification. International Journal of Remote Sensing, 23(4), 725-749. https://doi.org/10.1080/01431160110040323
Ma, L., Liu, Y., Zhang, X., Ye, Y., Yin, G., & Johnson, B. A. (2019). Deep learning in remote sensing applications: A meta-analysis and review. ISPRS Journal of Photogrammetry and Remote Sensing, 152, 166-177. https://doi.org/10.1016/j.isprsjprs.2019.04.015
Momeni, M., Saram, M. A., Latif, A. M., & Sheikhpour, R. (2020). Presenting a convolutional neural network based on dynamic adaptive fusion for noisy image classification. Signal and Data Processing, 46(17), 139-153.
Olofsson, P., Foody, G. M., Herold, M., Stehman, S. V., Woodcock, C. E., & Wulder, M. A. (2014). Good practices for estimating area and assessing the accuracy of land change. Remote Sensing of Environment, 148, 42-57. https://doi.org/10.1016/j.rse.2014.02.013
Rezaei, Y., Rezaei, A., Darke, F., & Azarafza, Z. (2021). Classification of polarimetric radar images based on a support vector machine and binary gravitational search algorithm. Signal and Data Processing, 47(18), 87-102.
Thanh Noi, P., & Kappas, M. (2018). Comparison of Random Forest, k-Nearest Neighbor, and Support Vector Machine classifiers for land cover classification using Sentinel-2 imagery. Sensors, 18(1), 18. https://doi.org/10.3390/s18010018
Tikuye, B. G., Rusnak, M., Manjunatha, B. R., & Jose, J. (2023). Land use and land cover change detection using the Random Forest approach: The case of the Upper Blue Nile River Basin, Ethiopia. Global Challenges, 7, 2300155. https://doi.org/10.1002/gch2.202300155
Turner, B. L., Lambin, E. F., & Reenberg, A. (2007). The emergence of land change science for global environmental change and sustainability. Proceedings of the National Academy of Sciences, 104(52), 20666-20671. https://doi.org/10.1073/pnas.0704119104
Yousefi, S., Tazeh, M., Mirzaee, S., Moradi, H. R., & Tavangar, S. (2011). Comparison of different classification algorithms of satellite images in preparing land use maps (Case study: Noor County). Journal of Remote Sensing and GIS in Natural Resources, 3(2), 15-26

Files

History

  • Receive Date: 02 February 2025
  • Revise Date: 19 April 2025
  • Accept Date: 25 May 2025
  • First Publish Date: 26 May 2025

Share

How to cite

Statistics