Comparison of boosting algorithms for land cover mapping using object-based classification
Submitted: 2025-08-27
|Accepted: 2025-12-02
|Published: 2025-12-17
Copyright (c) 2025 Ngoc Dung LUONG, Thanh Dong KHUC, Dinh Chieu VU
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Downloads
Keywords:
Land cover, boosting algorithms, simple non-iterative clustering, object-based classification, Sentinel-2 data
Supporting agencies:
Abstract:
Object-based image analysis (OBIA) is increasingly employed to enhance land cover classification accuracy from satellite imagery. This gap is addressed by systematically evaluating three leading boosting models - Gradient Boosting (GB), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting (LightGBM) for mapping six land cover classes in a complex mountainous landscape in Yen Bai province, Vietnam. We implement an OBIA using the Simple Non-Iterative Clustering (SNIC) algorithm for image segmentation on Sentinel-2 data. Spectral bands and derived indices (NDVI, NDBI, MBI) were used as predictive features. The findings demonstrate that all three models achieve high accuracy, with XGBoost emerging as the superior model, yielding an Overall Accuracy (OA) of 0.903 and a Kappa coefficient of 0.884. A key finding from the variable importance analysis is the differing feature reliance among algorithms: while GB and XGBoost prioritized the raw spectral information from the Red band (Band 4), LightGBM favored the derived Normalized Difference Built-up Index (NDBI). This research provides critical insights into the comparative strengths of boosting algorithms in an object-based framework, offering a valuable reference for selecting optimal models for high-resolution land cover monitoring initiatives.
References:
Achanta, R., & Süsstrunk, S. (2017). Superpixels and polygons using simple non-iterative clustering. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 4895–4904. https://doi.org/10.1109/CVPR.2017.520
Adam, H. E., Csaplovics, E., & Elhaja, M. E. (2016). A comparison of pixel-based and object-based approaches for land use land cover classification in semi-arid areas, Sudan. IOP Conference Series: Earth and Environmental Science, 37(1), 012061. https://doi.org/10.1088/1755-1315/37/1/012061
Aguirre-Gutiérrez, J., Seijmonsbergen, A. C., & Duivenvoorden, J. F. (2012). Optimizing land cover classification accuracy for change detection: A combined pixel-based and object-based approach in a mountainous area in Mexico. Applied Geography, 34, 29–37. https://doi.org/10.1016/j.apgeog.2011.10.010
Aryal, J., Sitaula, C., & Frery, A. C. (2023). Land use and land cover (LULC) performance modeling using machine learning algorithms: A case study of the city of Melbourne, Australia. Scientific Reports, 13(1), 13510. https://doi.org/10.1038/s41598-023-40564-0
Balha, A., Mallick, J. W., Pandey, S., Gupta, S., & Singh, C. K. (2021). A comparative analysis of different pixel and object-based classification algorithms using multi-source high spatial resolution satellite data for LULC mapping. Earth Science Informatics, 14(4), 2231–2247. https://doi.org/10.1007/s12145-021-00685-4
Ha, H., Bui, Q. D., Tran, D. T., Nguyen, D. Q., Bui, H. X., & Luu, C. (2024). Improving the forecast performance of landslide susceptibility mapping by using ensemble gradient boosting algorithms. Environment, Development and Sustainability. https://doi.org/10.1007/s10668-024-04694-3
Khuc, T. D., Luong, N. D., Dang, D. H., & Tran, V. A. (2025). Comparison of random forest and extreme gradient boosting algorithms in land cover classification in Van Yen District, Yen Bai Province, Vietnam. Journal of Hydro-Meteorology, 6(23), 50–59. https://doi.org/10.36335/VNJHM.2025(23).50-59
Khuc, T. D., Truong, X. Q., Tran, V. A., Bui, D. Q., Bui, D. P., Ha, H., Tran, T. H. M., Pham, T. T. T., & Yordanov, V. (2023). Comparison of multi-criteria decision making, statistics, and machine learning models for landslide susceptibility mapping in Van Yen District, Yen Bai Province, Vietnam. International Journal of Geoinformatics, 19(7), 33–45. https://doi.org/10.52939/ijg.v19i7.2743
M. Arpitha, Ahmed, S. A., & N, H. (2023). Land use and land cover classification using machine learning algorithms in Google Earth Engine. Earth Science Informatics, 16(4), 3057–3073. https://doi.org/10.1007/s12145-023-01073-w
Mousavinezhad, M., Feizi, A., & Aalipour, M. (2023). Performance evaluation of machine learning algorithms in change detection and change prediction of a watershed’s land use and land cover. International Journal of Environmental Research, 17(2), 29. https://doi.org/10.1007/s41742-023-00518-w
Nguyen, C. T., Chidthaisong, A., Kieu Diem, P., & Huo, L.-Z. (2021). A modified bare soil index to identify bare land features during agricultural fallow-period in Southeast Asia using Landsat 8. Land, 10(3), 231. https://doi.org/10.3390/land10030231
Saba, S. B., Ali, M., Turab, S. A., Waseem, M., & Faisal, S. (2023). Comparison of pixel, sub-pixel and object-based image analysis techniques for co-seismic landslides detection in seismically active area in Lesser Himalaya, Pakistan. Natural Hazards, 115(3), 2383–2398. https://doi.org/10.1007/s11069-022-05642-y
Shafizadeh-Moghadam, H., Khazaei, M., Alavipanah, S. K., & Weng, Q. (2021). Google Earth Engine for large-scale land use and land cover mapping: An object-based classification approach using spectral, textural and topographical factors. GIScience & Remote Sensing, 58(6), 914–928. https://doi.org/10.1080/15481603.2021.1947623
Tran, V. A., Khuc, T. D., Truong, X. Q., Nguyen, A. B., & Phi, T. T. (2024). Application of potential machine learning models in landslide susceptibility assessment: A case study of Van Yen District, Yen Bai Province, Vietnam. Quaternary Science Advances, 14, 100181. https://doi.org/10.1016/j.qsa.2024.100181
Truong, X. Q., Dang, N. H. D., Do, T. H., Tran, N. D., Do, T. T. N., Tran, V. A., Yordanov, V., Brovelli, M. A., & Khuc, T. D. (2023a). Random forest analysis of land use and land cover change using Sentinel-2 data in Van Yen, Yen Bai Province, Vietnam. In L. Q. Nguyen, L. K. Bui, X.-N. Bui, & H. T. Tran (Eds.), Advances in Geospatial Technology in Mining and Earth Sciences (pp. 429–445). Springer International Publishing.
Truong, X. Q., Tran, N. D., Dang, N. H. D., Do, T. H., Nguyen, Q. D., Yordanov, V., Brovelli, M. A., Duong, A. Q., & Khuc, T. D. (2023b). WebGIS and random forest model for assessing the impact of landslides in Van Yen District, Yen Bai Province, Vietnam (pp. 445–464). https://doi.org/10.1007/978-3-031-17808-5_27
Wei, X., Zhang, W., Zhang, Z., Huang, H., & Meng, L. (2023). Urban land use land cover classification based on GF-6 satellite imagery and multi-feature optimization. Geocarto International, 38(1), 2236579. https://doi.org/10.1080/10106049.2023.2236579
Zaabar, N., Niculescu, S., & Kamel, M. M. (2022). Application of convolutional neural networks with object-based image analysis for land cover and land use mapping in coastal areas: A case study in Ain Témouchent, Algeria. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 15, 5177–5189. https://doi.org/10.1109/JSTARS.2022.3185185
Zhang, W., He, Y., Wang, L., Liu, S., & Meng, X. (2023). Landslide susceptibility mapping using random forest and extreme gradient boosting: A case study of Fengjie, Chongqing. Geological Journal, 58(6), 2372–2387. https://doi.org/10.1002/gj.4683
Zhao, X. Y., Chen, J. X., Chen, G.-M., Xu, J. J., & Zhang, L. W. (2023). Prediction of ultimate condition of FRP-confined recycled aggregate concrete using a hybrid boosting model enriched with tabular generative adversarial networks. Thin-Walled Structures, 182, 110318. https://doi.org/10.1016/j.tws.2022.110318
