Comparative analysis of Random Forest and Support Vector Machine for LULC classification in Tashkent Region using Landsat-8 imagery

DOI: 10.35595/2414-9179-2025-1-31-519-532

View or download the article (Eng)

About the Authors

Nuriddin Mirjalalov

Mirzo Ulugbek National University of Uzbekistan,
4, University str., Tashkent, 100174, Republic of Uzbekistan,
E-mail: mirnur748@gmail.com

Nozimjon Teshaev

National Research University, Tashkent Institute of Irrigation and Agricultural Mechanization Engineers,
39, Koriy Niyoziy str., Tashkent, 100000, Republic of Uzbekistan,
E-mail: n.teshayev@tiiame.uz

Eshkabul Safarov

Mirzo Ulugbek National University of Uzbekistan,
4, University str., Tashkent, 100174, Republic of Uzbekistan,
E-mail: e.safarov@nuu.uz

Gerts Jasmina

Turin Polytechnic University,
17, Kichik Halka Yuli str., Tashkent, 100095, Republic of Uzbekistan,
E-mail: jasminagerts@tiiame.uz

Adbujalil Mominov

Alfraganus University,
2a, Yuqori Karakamish str., Tashkent, 100190, Republic of Uzbekistan,
E-mail: mirnur748@gmail.com

Anvar Pardaboyev

Tashkent State Technical University,
2b, University str., Tashkent, 100000, Republic of Uzbekistan,
E-mail: mirnur748@gmail.com

Abstract

Accurate Land Use/Land Cover (LULC) classification is essential for effective environmental monitoring, sustainable agricultural management, and informed urban planning. With increasing land transformation driven by urbanization, deforestation, and climate variability, reliable classification methods are needed to support data-driven decision-making. This study presents a comparative analysis of two widely used machine learning algorithms—Support Vector Machine (SVM) and Random Forest (RF)—for LULC classification in the Tashkent Region, Uzbekistan, using Landsat-8 Operational Land Imager (OLI) data for the June–July of 2024. The workflow involved critical preprocessing steps, cloud filtering, and band selection, followed by classification using six dominant LULC classes: water, bare land, built-up areas, cropland, snow, and forest. Ground truth samples were used to train and validate the models. Accuracy assessment was conducted using a confusion matrix, and performance was evaluated based on Overall Accuracy (OA) and the Kappa Coefficient (KC). The results revealed that the SVM classifier slightly outperformed RF, achieving an OA of 96.78 % and a KC of 0.96, compared to RF’s OA of 94.95 % and KC of 0.94. The superior performance of SVM is likely due to its effectiveness in handling high-dimensional feature spaces and separating non-linear class boundaries, especially in heterogeneous landscapes like Tashkent Region. While both algorithms showed strong potential, SVM demonstrated better precision in classifying cropland and forest areas. These findings highlight the importance of algorithm selection in remote sensing-based LULC studies. The study contributes to ongoing efforts to enhance land cover mapping accuracy using machine learning and offers valuable insights for land managers, urban developers, and environmental policymakers. Future research may consider the integration of multi-seasonal imagery, ancillary environmental data, and deep learning frameworks to further improve classification performance.

Keywords

land use/land cover (LULC), random forest (RF), support vector machine (SVM), Landsat-8, remote sensing, Tashkent Region

References

  1. Alikhanov B.S., Alikhanova R., Oymatov Z.F., Pulatov A. Land Cover Change in Tashkent Province during 1992–2018. IOP Conference Series: Materials Science and Engineering, 2020. V. 883. Art. 012088. DOI: 10.1088/1757-899X/883/1/012088.
  2. Alikhanov B., Pulatov B., Samiev L. The Detection of Past and Future Land Use and Land Cover Change in Ugam Chatkal National Park, Uzbekistan, Using CA-Markov and Random Forest Machine Learning Algorithms. Forum Geografi, 2024. V. 38. No. 2. P. 121–137. DOI: 10.23917/forgeo.v38i2.4221.
  3. Arpitha M., Sohrab A., Harishnaika N. Land Use and Land Cover Classification Using Machine Learning Algorithms in Google Earth Engine. Earth Science Informatics, 2023. V. 16. P. 3057–3073. DOI: 10.1007/s12145-023-01073-w.
  4. Aslanov I., Mukhtorov U., Mahsudov R., Makhmudova U., Alimova S., Djurayeva L., Ibragimov O. Applying Remote Sensing Techniques to Monitor Green Areas in Tashkent Uzbekistan. E3S Web of Conferences, 2021. V. 258. Art. 04012. DOI: 10.1051/e3sconf/202125804012.
  5. Aslanov I., Teshaev N., Jabbarov Z., Opp C., Oymatov R., Karimov Yu., Henebry G. Characterizing Land Surface Dynamics in Aral Sea Basin of Uzbekistan Using Climatic and Remote Sensing Data to Project Future Conditions. E3S Web of Conferences, 2024. V. 575. DOI: 10.1051/e3sconf/202457504009.
  6. Aslanov I., Teshaev N., Khayitov Kh., Mukhtorov U., Khaitbaeva J., Murodova D. Analysis of Desertification Trends in Central Asia Based on MODIS Data Using Google Earth Engine. E3S Web of Conferences, 2023. V. 443. Art. 06015. DOI: 10.1051/e3sconf/202344306015.
  7. Belgiu M., Drăguţ L. Random Forest in Remote Sensing: A Review of Applications and Future Directions. ISPRS Journal of Photogrammetry and Remote Sensing, 2016a. V. 114. P. 24–31. DOI: 10.1016/j.isprsjprs.2016.01.011.
  8. Bhungeni O., Ashadevi R., Gebreslasie M. Evaluating Machine-Learning Algorithms for Mapping LULC of the uMngeni Catchment Area, KwaZulu-Natal. Remote Sensing, 2024. V. 16. Art. 2219. DOI: 10.3390/rs16122219.
  9. Breiman L. Random Forests. Machine Learning, 2001. V. 45. Iss. 1. P. 5–32. DOI: 10.1023/A:1010933404324.
  10. Cortes C., Vapnik V. Support-Vector Networks. Machine Learning, 1995. V. 20. Iss. 3. P. 273–297. DOI: 10.1007/BF00994018.
  11. Dash S.S., Maity R. Effect of Climate Change on Soil Erosion Indicates a Dominance of Rainfall over LULC Changes. Journal of Hydrology: Regional Studies, 2023. V. 47. Art. 101373. DOI: 10.1016/j.ejrh.2023.101373.
  12. Elhadi A., Mutanga O., Rugege D. Multispectral and Hyperspectral Remote Sensing for Identification and Mapping of Wetland Vegetation: A Review. Wetlands Ecology and Management, 2010. V. 18. Iss. 3. P. 281–296. DOI: 10.1007/s11273-009-9169-z.
  13. Erdanaev E., Kappas M., Wyss D. Irrigated Crop Types Mapping in Tashkent Province of Uzbekistan with Remote Sensing-Based Classification Methods. Sensors, 2022. V. 22. Iss. 15. Art. 5863. DOI: 10.3390/s22155683.
  14. Farmonov N., Amankulova Kh., Khan Sh.N., Abdurakhimova M., Szatmári J., Khabiba T., Makhliyo R., Khodicha M., Mucsi L. Effectiveness of Machine Learning and Deep Learning Models at County-Level Soybean Yield Forecasting. Hungarian Geographical Bulletin, 2023. V. 72. No. 4. P. 383–398. DOI: 10.15201/hungeobull.72.4.4.
  15. Farmonov N., Esmaeili M., Abbasi-Moghadam D., Sharifi A., Amankulova Kh., Mucsi L. HypsLiDNet: 3D-2D CNN Model and Spatial-Spectral Morphological Attention for Crop Classification with DESIS and LiDAR Data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024. P. 1–34. DOI: 10.1109/JSTARS.2024.3418854.
  16. Gafforov Kh.Sh., Bao A., Rakhimov Sh., Liu T., Abdullaev F., Jiang L., Durdiev Kh., Duulatov E., Rakhimova M., Mukanov Ye. The Assessment of Climate Change on Rainfall-Runoff Erosivity in the Chirchik-Akhangaran Basin, Uzbekistan. Sustainability, 2020. V. 12. Iss. 8. Art. 3369. DOI: 10.3390/su12083369.
  17. Giles M.F. Status of Land Cover Classification Accuracy Assessment. Remote Sensing of Environment, 2002. V. 80. Iss. 1. P. 185–201. DOI: 10.1016/S0034-4257(01)00295-4.
  18. Gislason P.O., Benediktsson J.A., Sveinsson J.R. Random Forests for Land Cover Classification. Pattern Recognition Letters, Pattern Recognition in Remote Sensing (PRRS 2004), 2006. V. 27. Iss. 4. P. 294–300. DOI: 10.1016/j.patrec.2005.08.011.
  19. Huang C., Davis L.S., Townshend J.R.G. An Assessment of Support Vector Machines for Land Cover Classification. International Journal of Remote Sensing, 2002. V. 23. Iss. 4. P. 725–49. DOI: 10.1080/01431160110040323.
  20. Korhonen L., Hadi, Packalen P., Rautiainen M. Comparison of Sentinel-2 and Landsat-8 in the Estimation of Boreal Forest Canopy Cover and Leaf Area Index. Remote Sensing of Environment, 2017. V. 195. P. 259–74. DOI: 10.1016/j.rse.2017.03.021.
  21. López O.A.M., López A.M., Crossa J. Support Vector Machines and Support Vector Regression. Multivariate Statistical Machine Learning Methods for Genomic Prediction. Springer Nature, 2022. P. 337–378. DOI: 10.1007/978-3-030-89010-0_9.
  22. Maxwell A.E., Warner T.A., Fang F. Implementation of Machine-Learning Classification in Remote Sensing: An Applied Review. International Journal of Remote Sensing, 2018. V. 39. Iss. 9. P. 2784–2817. DOI: 10.1080/01431161.2018.1433343.
  23. Mirzaei A., Hossein B., Iman Kh. Enhancing Crop Classification Accuracy through Synthetic SAR-Optical Data Generation Using Deep Learning. ISPRS International Journal of Geo-Information, 2023. V. 12. Iss. 11. Art. 450. DOI: 10.3390/ijgi12110450.
  24. Mountrakis G., Im J., Ogole C. Support Vector Machines in Remote Sensing: A Review. ISPRS Journal of Photogrammetry and Remote Sensing, 2011. V. 66. Iss. 3. P. 247–259. DOI: 10.1016/j.isprsjprs.2010.11.001.
  25. Pal M. Random Forest Classifier for Remote Sensing Classification. International Journal of Remote Sensing, 2005. V. 26. Iss. 1. P. 217–222. DOI: 10.1080/01431160412331269698.
  26. Rodriguez-Galiano V.F., Ghimire B., Rogan J., Chica-Olmo M., Rigol-Sanchez J.P. An Assessment of the Effectiveness of a Random Forest Classifier for Land-Cover Classification. ISPRS Journal of Photogrammetry and Remote Sensing, 2012a. V. 67. P. 93–104. DOI: 10.1016/j.isprsjprs.2011.11.002.
  27. Sajib Md.Q.U., Wang T. Estimation of Land Surface Temperature in an Agricultural Region of Bangladesh from Landsat-8: Intercomparison of Four Algorithms. Sensors, 2020. V. 20. Iss. 6. Art. 1778. DOI: 10.3390/s20061778.
  28. Schmidt M., Pringle M., Devadas R., Denham R., Tindall D. A Framework for Large-Area Mapping of Past and Present Cropping Activity Using Seasonal Landsat Images and Time Series Metrics. Remote Sensing, 2016. V. 8. Iss. 4. Art. 312. DOI: 10.3390/rs8040312.
  29. Sharipov Sh., Khayitmurodov A. The Impacts of Green Spaces on Mitigating the Urban Hot Island Effect in the City of Tashkent. BIO Web of Conferences, 2024. V. 105. Art. 06013. DOI: 10.1051/bioconf/202410506013.
  30. Sultan M., Saleous N., Issa S., Dahy B., Sami M. Optimizing Land Use Classification Using Google Earth Engine: A Comparative Analysis of Machine Learning Algorithms. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2025. V. X-G-2025. P. 863–869. DOI: 10.5194/isprs-annals-X-G-2025-863-2025.
  31. Teshaev N., Mamadaliyev B., Ibragimov A., Khasanov S. The Soil-Adjusted Vegetation Index for Soil Salinity Assessment in Uzbekistan. InterCarto. InterGIS. Proceedings of the International Conference. Moscow: Lomonosov Moscow State University, Faculty of Geography, 2020. V. 26. Part 3. P. 324–333. DOI: 10.35595/2414-9179-2020-3-26-324-333.
  32. Tian Sh., Zhang X., Tian J., Sun Q. Random Forest Classification of Wetland Landcovers from Multi-Sensor Data in the Arid Region of Xinjiang, China. Remote Sensing, 2016. V. 8. Iss. 11. Art. 954. DOI: 10.3390/rs8110954.
  33. Vapnik V.N. Introduction: Four Periods in the Research of the Learning Problem. The Nature of Statistical Learning Theory. Springer Nature, 2000. P. 1–15. DOI: 10.1007/978-1-4757-3264-1_1.
  34. Wei P., Ye H., Qiao Sh., Liu R., Nie Ch., Zhang B., Song L., Huang Sh. Early Crop Mapping Based on Sentinel-2 Time-Series Data and the Random Forest Algorithm. Remote Sensing, 2023. V. 15. Iss. 13. Art. 3212. DOI: 10.3390/rs15133212.
  35. Zhang Ch., Dong J., Xie Ya., Zhang X., Ge Q. Mapping Irrigated Croplands in China Using a Synergetic Training Sample Generating Method, Machine Learning Classifier, and Google Earth Engine. International Journal of Applied Earth Observation and Geoinformation, 2022. V. 112. Art. 102888. DOI: 10.1016/j.jag.2022.102888.
  36. Zhou Ya., Jiang H., Wang Zh., Yang X., Geng E. Assessment of Four Typical Topographic Corrections in Landsat TM Data for Snow Cover Areas. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2016. V. XLI-B1. P. 157–162. DOI: 10.5194/isprs-archives-XLI-B1-157-2016.

For citation: Mirjalalov N., Teshaev N., Safarov E., Jasmina G., Mominov A., Pardaboyev A. Comparative analysis of Random Forest and Support Vector Machine for LULC classification in Tashkent Region using Landsat-8 imagery. InterCarto. InterGIS. Moscow: MSU, Faculty of Geography, 2025. V. 31. Part 1. P. 519–532. DOI: 10.35595/2414-9179-2025-1-31-519-532