Global-scale explainable AI assessment for OBIA-based classification using Deep Learning and Machine Learning methods

• Autorzy: Kavzoglu Taskin , Uzun Yakup Kaan , Berkan Emre , Yilmaz Elif Ozlem

Identyfikatory

DOI

10.24425/agg.2025.150692

• Języki publikacji: EN

Abstrakty

EN

Over the past decade, object-based image analysis (OBIA) has gained prominence as a widely adopted method for generating land use/land cover (LULC) maps. This study aims to evaluate the performance of various classification algorithms within the OBIA framework using SPOT-6 satellite imagery. The research methodology involved segmenting the images with the multi-resolution segmentation (MRS) algorithm, followed by the application of convolutional neural networks (CNN), random forest (RF), and support vector machine (SVM) algorithms for classification. The study was conducted in the Perpignan province, located in the Pyrénées-Orientales region of France. After the segmentation stage, CNN, RF, and SVM classifiers were employed to classify the image segments based on both spectral and spatial attributes. The accuracy of the resulting thematic maps was assessed using standard metrics, including overall accuracy (OA), the Kappa coefficient (KC), and the F��score (FS). Of the three classifiers, CNN achieved the highest overall accuracy at 91.28%, outperforming SVM, which attained an OA of 90.50%, and RF, which recorded an OA of 87.28%. Additionally, this study explored the integration of explainable artificial intelligence (AI) techniques, specifically the Shapley Additive Explanations (SHAP) algorithm, to enhance the interpretability of the machine learning models. This approach fosters greater trust, accountability, and acceptance in decision-making processes. By leveraging SHAP values, the study provides deeper insights into the decision-making processes of the CNN, SVM, and RF classifiers, ultimately enhancing the transparency and comprehensibility of these models.

Słowa kluczowe

EN

support vector machine; convolutional neural network; random forest; XAI; object-based image analysis

PL

użytkowanie gruntów; las sosnowy; obraz satelitarny; sieć neuronowa

• Wydawca: Komitet Geodezji PAN
• Czasopismo: Advances in Geodesy and Geoinformation
• Rocznik: 2025
• Tom: Vol. 74, no. 1
• Strony: art. no. e62, 2025
• Opis fizyczny: Bibliogr. 48 poz., rys., tab., wykr.

Twórcy

autor

Kavzoglu Taskin

• Gebze Technical University, Kocaeli, Turkey

• https://orcid.org/0000-0002-9779-3443

autor

Uzun Yakup Kaan

• Gebze Technical University, Kocaeli, Turkey

• https://orcid.org/0009-0007-2556-8363

autor

Berkan Emre

• Gebze Technical University, Kocaeli, Turkey

• https://orcid.org/0009-0009-9861-6113

autor

Yilmaz Elif Ozlem

• Gebze Technical University, Kocaeli, Turkey

• https://orcid.org/0000-0002-6853-2148

Bibliografia

• 1. Ameslek, O., Zahir, H., Latifi, H. et al. (2024). Combining OBIA, CNN, and UAV imagery for automated detection and mapping of individual olive trees. Smart Agric. Technol., 9, 100546. DOI:10.1016/j.atech.2024.100546.
• 2. Arfa, A., and Minaei, M. (2024). Utilizing multitemporal indices and spectral bands of Sentinel-2 to enhance land use and land cover classification with Random Forest and Support Vector Machine. Adv. Space Res., 74(11), 5580–5590. DOI: 10.1016/j.asr.2024.08.062.
• 3. Azeez, O.S., Shafri, H.Z., Alias, A.H. et al. (2022). Integration of Object-Based Image Analysis and Convolutional Neural Network for the classification of high-resolution satellite image: a comparative assessment. Appl. Sci., 12(21), 10890. DOI: 10.3390/app122110890.
• 4. Belgiu, M., and Dragut, L. (2016). Random forest in remote sensing: a review of applications and future directions. ISPRS J. Photogramm. Remote Sens., 114, 24–31. DOI: 10.1016/j.isprsjprs.2016.01.011.
• 5. Blaschke, T. (2010). Object based image analysis for remote sensing. ISPRS J. Photogramm. and Remote Sens., 65(1), 2–16. DOI: 10.1016/j.isprsjprs.2009.06.004.
• 6. Breiman, L. (2001). Random Forests. Mach. Learn., 45, 5–32. DOI: 10.1023/A:1010933404324.
• 7. Chowdhury, M.S. (2024). Comparison of accuracy and reliability of Random Forest, Support Vector Machine, Artificial Neural Network and Maximum Likelihood method in land use/cover classification of urban setting. Environ. Chall., 14, 100800. DOI: 10.1016/j.envc.2023.100800.
• 8. Chughtai, A.H., Abbasi, H., and Karas, I.R. (2021). A review on change detection method and accuracy assessment for land use land cover. Remote Sens. Appl.: Soc. Environ., 22, 100482. DOI:10.1016/j.rsase.2021.100482.
• 9. Colkesen, I., and Kavzoglu, T. (2018). Selection of optimal object features in object-based image analysis using filter-based algorithms. J. Indian Soc. Remote Sens., 46(8), 1233-1242. DOI: 10.1007/s12524-018-0807-x.
• 10. Colkesen, I., and Kavzoglu, T. (2019). Comparative evaluation of decision-forest algorithms for object-based land use and land cover mapping. In H.R. Pourghasemi and C. Gokceoglu (Eds.), Spatial Modeling in GIS and R for Earth and Environmental Sciences, 499-517. DOI: 10.1016/B978-0-12-815226-3.00023-5.
• 11. Cortes, C., and Vapnik, V. (1995). Support-Vector Networks. Mach. Learn., 20. 273–297.
• 12. Dragut, L., Csillik, O., Eisank, C. et al. (2014). Automated parameterisation for multi-scale image segmentation on multiple layers. ISPRS J. Photogramm. Remote Sens. 88, 119–127. DOI:10.1016/j.isprsjprs.2013.11.018.
• 13. Foody, G.M. (2020). Explaining the unsuitability of the Kappa coefficient in the assessment and comparison of the accuracy of thematic maps obtained by image classification. Remote Sens. Environ., 239, 111630. DOI: 10.1016/j.rse.2019.111630.
• 14. Franklin, S.E., and Ahmed, O.S. (2017). Object-based wetland characterization using Radarsat-2 Quad-Polarimetric SAR Data, Landsat-8 OLI imagery, and airborne lidar-derived geomorphometric Variables. Photogramm. Eng. Rem. S., 83(1), 27–36. DOI: 10.14358/PERS.83.1.27.
• 15. Gamanya, R., De Maeyer, P., and De Dapper, M. (2009). Object-oriented change detection for the city of Harare, Zimbabwe. Expert Systems with Applications, 36(1), 571–588. DOI:10.1016/j.eswa.2007.09.067.
• 16. Hossain, M.D., and Chen, D. (2019). Segmentation for Object-Based Image Analysis (OBIA): A review of algorithms and challenges from remote sensing perspective. ISPRS J. of Photogramm. and Remote Sens., 150, 115–134. DOI: 10.1016/j.isprsjprs.2019.02.009.
• 17. Kattenborn, T., Leitloff, J., Schiefer, F. et al. (2021). Review on Convolutional Neural Networks (CNN) in vegetation remote sensing. ISPRS J. Photogramm. Remote Sens., 173, 24–49. DOI:10.1016/j.isprsjprs.2020.12.010.
• 18. Kavzoglu, T. (2017). Object-oriented random forest for high resolution land cover mapping using quickbird-2 imagery. In P. Samui, S. Sekhar, and V. E. Balas (Eds.), Handbook of Neural Computation, 607–619. Academic Press. DOI: 10.1016/B978-0-12-811318-9.00033-8.
• 19. Kavzoglu, T., and Tonbul, H. (2017). A Comparative Study of Segmentation Quality for Multi-Resolution Segmentation and Watershed Transform, Istanbul, Turkey. 8th International Conference on Recent Advances in Space Technologies (RAST), 19–22 June (pp. 113–117), Istanbul, Turkey. DOI:10.1109/RAST.2017.8002984.
• 20. Kavzoglu, T., and Tonbul, H. (2018). An experimental comparison of multi-resolution segmentation, SLIC And K-Means Clustering for Object-Based Classification of VHR Imagery. Int. J. Remote Sens., 39(18). 6020–6036. DOI: 10.1080/01431161.2018.1506592.
• 21. Kavzoglu, T., Tonbul, H., Yildiz Erdemir, M. et al. (2018). Dimensionality Reduction and Classification of Hyperspectral Images Using Object-Based Image Analysis. J. Indian Soc. Remote Sens., 46, 1297–1306. DOI: 10.1007/s12524-018-0803-1.
• 22. Kavzoglu, T., Teke, A., and Yilmaz, E.O. (2021). Shared blocks-based ensemble deep learning for shallow landslide susceptibility mapping. Remote Sens., 13(23), 4776. DOI: 10.3390/rs13234776.
• 23. Kavzoglu, T., and Teke, A. (2022). Predictive performances of ensemble machine learning algorithms in landslide susceptibility mapping using Random Forest, Extreme Gradient Boosting (Xgboost) and Natural Gradient Boosting (NGBoost). Arab. J. Sci. Eng., 47(6), 7367–7385. DOI: 10.1007/s13369-022-06560-8.
• 24. Kavzoglu, T., and Yilmaz, E.Ö. (2022). Analysis of patch and sample size effects for 2D-3D CNN models using multiplatform dataset: hyperspectral image classification of ROSIS and Jilin-1 GP01 imagery. Turk. J. Elec. Eng. Comp. Sci., 30(6), 2124–2144. DOI: 10.55730/1300-0632.3929.
• 25. Kavzoglu, T., and Bilucan, F. (2023). Effects of auxiliary and ancillary data on LULC classification in a heterogeneous environment using optimized random forest algorithm. Earth Sci. Inform., 16(1), 415–435. DOI: 10.1007/s12145-022-00874-9.
• 26. Kavzoglu, T., Tso, B., and Mather, P.M. (2024). Classification Methods for Remotely Sensed Data. Third Edition, Boca Raton: CRC Press. DOI: 10.1201/9781003439172.
• 27. Linhui, L., Weipeng, J., and Huihui, W. (2020). Extracting the forest type from remote sensing images by Random Forest. IEEE Sens. J., 21(16), 17447–17454. DOI: 10.1109/JSEN.2020.3045501.
• 28. Lundberg, S.M., and Lee., S.I. (2017). A unified approach to interpreting model predictions. In Proceedings of NIPS’17, December 4–9, Long Beach, CA, USA, 4765-4774.
• 29. Lundberg, S.M., Erion, G., Chen, H. et al. (2020). From Local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell., 2(1), 56–67. DOI: 10.1038/s42256-019-0138-9.
• 30. Mantero, P., Moser, G. and Serpico, S.B. (2005). Partially supervised classification of remote sensing images through SVM-based probability density estimation. IEEE Trans. Geosci. Remote Sens., 43, 559–570. DOI: 10.1109/TGRS.2004.842022.
• 31. Moharram, M.A., and Sundaram, D.M. (2023). Land use and land cover classification with hyperspectral data: a comprehensive review of methods, challenges and future directions. Neurocomputing, 536, 90–113. DOI: 10.1016/j.neucom.2023.03.025.
• 32. Mountrakis, G., Im, J., and Ogole, C. (2011). Support vector machines in remote sensing: A review. ISPRS J. Photogramm. Remote Sens., 66(3), 247–259. DOI: 10.1016/j.isprsjprs.2010.11.001.
• 33. Pal, M. (2005). Random Forest classifier for remote sensing classification. Int. J. of Remote Sens., 26(1), 217–222. DOI: 10.1080/01431160412331269698.
• 34. Pal, M., and Mather, P.M. (2005). Support Vector Machines for classification in remote sensing. Int. J. Remote Sens., 26(5), 1007–1011. DOI: 10.1080/01431160512331314083.
• 35. Pal, M., and Foody, G.M. (2012). Evaluation of SVM, RVM and SMLR for accurate image classification with limited ground data. IEEE J. Sel. Top. Appl. Earth Obs., 5, 1344–1355. DOI: 10.1109/JSTARS. 2012.2215310.
• 36. Panda, K.C., Singh, R.M., and Singh, S.K. (2024). Advanced CMD predictor screening approach coupled with cellular automata-artificial neural network algorithm for efficient land use-land cover change prediction. J. Cleaner Prod., 449, 141822. DOI: 10.1016/j.jclepro.2024.141822.
• 37. Pandey, P.C., Koutsias, N., Petropoulos, G.P. et al. (2021). Land use/land cover in view of earth observation: data sources, input dimensions, and classifiers – a review of the state of the art. Geocarto Int., 36(9), 957–988. DOI: 10.1080/10106049.2019.1629647.
• 38. Sheykhmousa, M., Mahdianpari, M., Ghanbari, H. et al. (2020). Support Vector Machine versus Random Forest for remote sensing image classification: a meta-analysis and systematic review. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 13, 6308–6325. DOI: 10.1109/JSTARS.2020.3026724.
• 39. Singh, S.K., Srivastava, P.K., Szabó, S. et al. (2017). Landscape transform and spatial metrics for mapping spatiotemporal land cover dynamics using earth observation data-sets. Geocarto Int., 32(2), 113–127. DOI: 10.1080/10106049.2015.1130084.
• 40. Song, J., Gao, S., Zhu, Y. et al. (2019). A survey of remote sensing image classification based on CNNs. Big Earth Data, 3(3), 232–254. DOI: 10.1080/20964471.2019.1657720.
• 41. Tonbul, H., and Kavzoglu, T. (2020a). Semi-Automatic building extraction from Worldview-2 imagery using Taguchi optimization. Photogramm. Eng. Remote Sensing, 86(9), 547–555. DOI:10.14358/PERS.86.9.547.
• 42. Tonbul, H., and Kavzoglu, T. (2020b). A spectral band based comparison of unsupervised segmentation evaluation methods for image segmentation parameter optimization. Int. J. Environ. Geoinformatics, 7(2), 132–139. DOI: 10.30897/ijegeo.641216.
• 43. Tonbul, H., Colkesen, I., and Kavzoglu, T. (2022). Pixel-and object-based ensemble learning for forest burn severity using USGS FIREMON and Mediterranean condition dNBRs in Aegean ecosystem (Turkey). Adv. Space Res., 69(10), 3609–3632. DOI: 10.1016/j.asr.2022.02.051.
• 44. Townshend, J.R.G. (1992). Land cover. Int. J. Remote Sensing, 13(6–7), 1319–1328. DOI:10.1080/01431169208904193.
• 45. Vizzari, M. (2022). PlanetScope, Sentinel-2, and Sentinel-1 data integration for object-based land cover classification in Google Earth Engine. Remote Sens., 14(11), 2628. DOI: 10.3390/rs14112628.
• 46. Wang, Y., Sun, Y., Cao, X. et al. (2023). A review of regional and global scale Land Use/Land Cover (LULC) mapping products generated from satellite remote sensing. ISPRS J. Photogramm. Remote Sens., 206, 311–334. DOI: 10.1016/j.isprsjprs.2023.11.014.
• 47. Yilmaz, E.O., and Kavzoglu, T. (2024). Quality assessment for multi-resolution segmentation and segmentanything model using WorldView-3 imagery. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLVIII-4/W9-2024, 48, 383–390. DOI: 10.5194/isprs-archives-XLVIII-4-W9-2024-383-2024.
• 48. Zhang, X., Chen, L., and Jia, X. (2018). Deep learning for remote sensing image classification: A comprehensive review. IEEE Geosci. Remote Sens Mag., 6(3), 6–20. DOI: 10.1002/widm.1264.

Uwagi

Opracowanie rekordu ze środków MNiSW, umowa nr POPUL/SP/0154/2024/02 w ramach programu "Społeczna odpowiedzialność nauki II" - moduł: Popularyzacja nauki (2025)