Design and Evaluation of an Optimized Random Forest Classification Framework for Tropical Forest Degradation Detection
DOI:
https://doi.org/10.30871/jaic.v10i2.12494Keywords:
Random Forest, Tropical forest monitoring, Automated classification, Change detection, Deforestation, Machine learning, Landsat 8, Congo BasinAbstract
Tropical forest degradation monitoring in resource-constrained environments requires classification approaches that balance predictive accuracy with computational efficiency. While deep learning methods often achieve high classification performance, their reliance on large labeled datasets and GPU-based infrastructure limits operational deployment in many conservation contexts. This study proposes and evaluates a systematically optimized Random Forest classification framework for tropical forest degradation detection in the Congo Basin using Landsat 8 multispectral imagery. A multi-class model distinguishing intact forest, degraded forest, and non-forest was developed using six spectral bands and four ecologically motivated vegetation indices (NDVI, EVI, NDMI, and NBR). Ground truth data consisted of 450 stratified samples (180 intact forest, 150 degraded forest, 120 non-forest; imbalance ratio 1.5) validated through field surveys and high-resolution imagery collected during the 2022 dry season. Class imbalance was mitigated through balanced class weighting during training. Model optimization was achieved through controlled hyperparameter tuning with five-fold cross-validation, targeting improved generalization while preserving computational efficiency. The optimized Random Forest achieved an overall accuracy of 89.3% (Kappa = 0.834) on an independent test set, significantly outperforming CART and SVM baselines while maintaining a training time under two minutes on standard CPU hardware. Class-specific F1-scores ranged from 0.88 to 0.91, indicating balanced performance across all land-cover classes. Processing the full 7500 km² study area required approximately 14 minutes without GPU acceleration. These results demonstrate that systematically optimized ensemble learning provides a robust, interpretable, and operationally deployable solution for tropical forest degradation monitoring, particularly suited to conservation agencies operating under limited computational and financial resources.
Downloads
References
[1] Asner, G. P. (2001). Cloud cover in Landsat observations of the Brazilian Amazon. International Journal of Remote Sensing, 22(18), 3855–3862.
[2] Belgiu, M., & Drăguţ, L. (2016). Random forest in remote sensing: A review of applications and future directions. ISPRS Journal of Photogrammetry and Remote Sensing, 114, 24–31.
[3] Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.
[4] Campos-Taberner, M., García-Haro, F. J., Martínez, B., et al. (2020). Understanding deep learning in land use classification based on Sentinel-2 time series. Scientific Reports, 10(1), 17188.
[5] Chavez, P. S. (1996). Image-based atmospheric corrections Revisited and improved. Photogrammetric Engineering and Remote Sensing, 62(9), 1025–1036.
[6] Dargie, G. C., Lewis, S. L., Lawson, I. T., et al. (2017). Age, extent and carbon storage of the central Congo Basin peatland complex. Nature, 542(7639), 86–90.
[7] De Wasseige, C., de Marcken, P., Bayol, N., et al. (2012). Les forêts du Bassin du Congo État des forêts 2010. Office des publications de l'Union européenne, Luxembourg.
[8] FAO. (2020). Global Forest Resources Assessment 2020. Food and Agriculture Organization of the United Nations, Rome.
[9] Foody, G. M. (2002). Status of land cover classification accuracy assessment. Remote Sensing of Environment, 80(1), 185–201.
[10] Friedl, M. A., & Brodley, C. E. (1997). Decision tree classification of land cover from remotely sensed data. Remote Sensing of Environment, 61(3), 399–409.
[11] Fuller, D. O. (2006). Tropical forest monitoring and remote sensing. Singapore Journal of Tropical Geography, 27(1), 15–29.
[12] Hammer, D., Kraft, R., & Wheeler, D. (2014). FORMA: Forest Monitoring for Action. Center for Global Development Working Paper 192.
[13] Hansen, M. C., Potapov, P. V., Moore, R., et al. (2013). High-resolution global maps of 21st-century forest cover change. Science, 342(6160), 850–853.
[14] Herold, M., & Skutsch, M. (2011). Monitoring, reporting and verification for national REDD+ programmes. Environmental Research Letters, 6(1), 014002.
[15] ICCN. (2012). Plan de gestion de la Réserve Naturelle de Tumba-Lediima 2012–2016. Institut Congolais pour la Conservation de la Nature, Kinshasa.
[16] Japkowicz, N., & Stephen, S. (2002). The class imbalance problem: A systematic study. Intelligent Data Analysis, 6(5), 429–449.
[17] Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159–174.
[18] Lewis, S. L., Lopez-Gonzalez, G., Sonké, B., et al. (2009). Increasing carbon storage in intact African tropical forests. Nature, 457(7232), 1003–1006.
[19] Ma, L., Liu, Y., Zhang, X., et al. (2019). Deep learning in remote sensing applications: A meta-analysis and review. ISPRS Journal of Photogrammetry and Remote Sensing, 152, 166–177.
[20] Maxwell, A. E., Warner, T. A., & Fang, F. (2018). Implementation of machine-learning classification in remote sensing: An applied review. International Journal of Remote Sensing, 39(9), 2784–2817.
[21] Mountrakis, G., Im, J., & Ogole, C. (2011). Support vector machines in remote sensing: A review. ISPRS Journal of Photogrammetry and Remote Sensing, 66(3), 247–259.
[22] Muteya, H. K., et al. (2022). Monitoring deforestation dynamics in the Congo Basin using satellite remote sensing. Environmental Monitoring and Assessment, 194(2), 112.
[23] Potapov, P. V., Turubanova, S. A., Hansen, M. C., et al. (2012). Quantifying forest cover loss in Democratic Republic of the Congo, 2000–2010, with Landsat ETM+ data. Remote Sensing of Environment, 122, 106–116.
[24] Reiche, J., Hamunyela, E., Verbesselt, J., et al. (2018). Improving near-real time deforestation monitoring in tropical dry forests by combining dense Sentinel-1 time series with Landsat and ALOS-2. Remote Sensing of Environment, 204, 147–161.
[25] Roberts, D. R., Bahn, V., Ciuti, S., et al. (2017). Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. Ecography, 40(8), 913–929.
[26] Rodriguez-Galiano, V. F., Ghimire, B., Rogan, J., et al. (2012). An assessment of the effectiveness of a random forest classifier for land-cover classification. ISPRS Journal of Photogrammetry and Remote Sensing, 67, 93–104.
[27] Shapiro, A. C., Grantham, H. S., Aguilar-Amuchastegui, N., et al. (2016). Forest condition in the Congo Basin for the assessment of ecosystem conservation status. Ecological Indicators, 60, 1074–1085.
[28] Souza, C. M., Roberts, D. A., & Cochrane, M. A. (2013). Combining spectral and spatial information to map canopy damage from selective logging. Remote Sensing of Environment, 98(2–3), 329–343.
[29] Tucker, C. J. (1979). Red and infrared linear combinations for monitoring vegetation. Remote Sensing of Environment, 8(2), 127–150.
[30] Tyukavina, A., Hansen, M. C., Potapov, P., et al. (2018). Congo Basin forest loss dominated by increasing smallholder clearing. Science Advances, 4(11), eaat2993.
[31] Woodcock, C. E., Allen, R., Anderson, M., et al. (2008). Free access to Landsat imagery. Science, 320(5879), 1011.
[32] Zhang, C., Sargent, I., Pan, X., et al. (2021). Joint deep learning for land cover and land use classification. Remote Sensing of Environment, 221, 173–187.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Frey Sylvestre, Simboni Simboni Tege, Mabela Makengo Ronstand, Mbuyi Mukendi Eugene, Anzola Kibamba Nestor, Kafunda Katalay Pierre
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
