Research Article
Classification of Urban Vegetation Utilizing Spectral Indices and DEM with Ensemble Machine Learning Methods
Abstract
Detection and monitoring of urban vegetation is a subject in many sustainable development goal studies. Detection of green areas and their decreasing rates with increasing urbanization are followed with interest by municipalities and planners. Due to their high cost-effectiveness, unmanned aerial vehicles (UAVs) have been used extensively in agriculture and forest management. In this study, low vegetation, tree and non-vegetation classes were classified using ensemble machine learning methods on a university campus. After the pre-processing steps of the images obtained via UAV, datasets consisting of different bands and indices were created for classification and the effect of the DEM layer was also investigated. Four machine learning classifiers were implemented, namely XGBoost, LightGBM, Gradient Boosting and CatBoost. According to the results, the highest classification performances are achieved when vegetation indices and DEM are used together. The CatBoost method obtained 90.2% accuracy and 86.9% F1-score. It is understood that the classification of multispectral aerial images with ML has shown promising results in the detection of vegetation in urban areas
References
- Alpaydin, E. (2020). Introduction to machine learning. MIT press.
- Atik, M. E., & Arkali, M. (2025). Comparative Assessment of the Effect of Positioning Techniques and Ground Control Point Distribution Models on the Accuracy of UAV-Based Photogrammetric Production. Drones, 9(1), 15. https://doi.org/10.3390/drones9010015
- Atik, M. E., Duran, Z., & Seker, D. Z. (2024). Explainable Artificial Intelligence for Machine Learning-Based Photogrammetric Point Cloud Classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.
- Atik, Ş. Ö. (2023). Çok Yüksek Çözünürlüklü Uydu Görüntülerinden Bina Çıkarımında Derin Öğrenme ve Çoklu-Çözünürlüklü Bölütleme Kullanılarak Nesne-Tabanlı Entegrasyon. Türkiye Uzaktan Algılama Dergisi, 5(2), 67-77.
- Atik, S. O., & Atik, M. E. (2024). Optimal band selection using explainable artificial intelligence for machine learning-based hyperspectral image classification. Journal of Applied Remote Sensing, 18(4), 042604-042604.
- Barnes, E. M., Clarke, T. R., Richards, S. E., Colaizzi, P. D., Haberland, J., Kostrzewski, M., ... & Moran, M. S. (2000, July). Coincident detection of crop water stress, nitrogen status and canopy density using ground based multispectral data. In Proceedings of the fifth international conference on precision agriculture, Bloomington, MN, USA (Vol. 1619, No. 6).
- Bayirhan, İ., & Gazioğlu, C. (2020). Use of Unmanned Aerial Vehicles (UAV) and Marine Environment Simulator in Oil Pollution Investigations. Baltic Journal of Modern Computing, 8(2).
- Cao, Q., Li, M., Yang, G., Tao, Q., Luo, Y., Wang, R., & Chen, P. (2024). Urban Vegetation Classification for Unmanned Aerial Vehicle Remote Sensing Combining Feature Engineering and Improved DeepLabV3+. Forests, 15(2), 382. https://doi.org/10.3390/f15020382
- Chen, T., & Guestrin, C. (2016, August). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining (pp. 785-794). Friedman, J. H. (2002). Stochastic gradient boosting. Computational statistics & data analysis, 38(4), 367-378.
- Gallardo-Salazar, J. L., & Pompa-García, M. (2020). Detecting individual tree attributes and multispectral indices using unmanned aerial vehicles: Applications in a pine clonal orchard. Remote Sensing, 12(24), 4144.
- Gazioğlu, C., Varol, Ö. E., Şeker, D. Z., & Çağlar, N. (2017, December). Determination of the Environmental Impacts of Marine Accidents Using UAV and RS Technologies. In 19th MESAEP Symposium on Environmental and Health Inequity.
- Gitelson, A.A., Kaufman, Y.J., & Merzlyak, M.N. (1996). Use of a green channel in remote sensing of global vegetation from EOSMODIS. Remote Sens Environ, 58, 289-298.
- Hartling, S., Sagan, V., & Maimaitijiang, M. (2021). Urban tree species classification using UAV-based multi-sensor data fusion and machine learning. GIScience & Remote Sensing, 58(8), 1250–1275. https://doi.org/10.1080/15481603.2021.1974275
- Hu, K., Liu, J., Xiao, H., Zeng, Q., Liu, J., Zhang, L., ... & Wang, Z. (2024). A new BWO-based RGB vegetation index and ensemble learning strategy for the pests and diseases monitoring of CCB trees using unmanned aerial vehicle. Frontiers in Plant Science, 15, 1464723.
- Iglhaut, J., Cabo, C., Puliti, S., Piermattei, L., O’Connor, J., & Rosette, J. (2019). Structure from motion photogrammetry in forestry: A review. Current Forestry Reports, 5, 155-168.
- Kaya, Z., & Dervisoglu, A. (2023). Determination of urban areas using Google Earth Engine and spectral indices; Esenyurt case study. International Journal of Environment and Geoinformatics, 10(1), 1-8.
- Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., ... & Liu, T. Y. (2017). Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems, 30.
- Liu, S., Jin, X., Bai, Y., Wu, W., Cui, N., Cheng, M., ... & Yin, D. (2023). UAV multispectral images for accurate estimation of the maize LAI considering the effect of soil background. International Journal of Applied Earth Observation and Geoinformation, 121, 103383.
- López-García, P., Intrigliolo, D., Moreno, M. A., Martínez-Moreno, A., Ortega, J. F., Pérez-Álvarez, E. P., & Ballesteros, R. (2022). Machine Learning-Based Processing of Multispectral and RGB UAV Imagery for the Multitemporal Monitoring of Vineyard Water Status. Agronomy, 12(9), 2122. https://doi.org/10.3390/agronomy12092122
- Lowe, D. G. (2004). Distinctive image features from scale-invariant keypoints. International journal of computer vision, 60, 91-110.
- Momm, H., & Easson, G. (2011, April). Feature extraction from high-resolution remotely sensed imagery using evolutionary computation. In Evolutionary Algorithms. IntechOpen.
- Munyati, C. (2000). Wetland change detection on the Kafue Flats, Zambia, by classification of a multitemporal remote sensing image dataset. International journal of remote sensing, 21(9), 1787-1806.
- Öztürk, M. Y., & Çölkesen, İ. (2021). The impacts of vegetation indices from UAV-based RGB imagery on land cover classification using ensemble learning. Mersin Photogrammetry Journal, 3(2), 41-47.
- Peñuelas, J., Filella, I., Biel, C., Serrano, L., & Save, R. (1993). The reflectance at the 950–970 nm region as an indicator of plant water status. International journal of remote sensing, 14(10), 1887-1905.
- Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (2018). CatBoost: unbiased boosting with categorical features. Advances in neural information processing systems, 31.
- Pu, R., Gong, P., & Yu, Q. (2008). Comparative analysis of EO-1 ALI and Hyperion, and Landsat ETM+ data for mapping forest crown closure and leaf area index. Sensors, 8(6), 3744-3766.
- Rondeaux, G., Steven, M., & Baret, F. (1996). Optimization of soil-adjusted vegetation indices. Remote sensing of environment, 55(2), 95-107. Schonberger, J. L., & Frahm, J. M. (2016). Structure-from-motion revisited. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 4104-4113).
- Sefercik, U. G., Nazar, M., Aydin, I., Büyüksalih, G., Gazioglu, C., & Bayirhan, I. (2024). Comparative analyses for determining shallow water bathymetry potential of multispectral UAVs: case study in Tavşan Island, Sea of Marmara. Frontiers in Marine Science, 11, 1388704.
- Shu, M., Fei, S., Zhang, B., Yang, X., Guo, Y., Li, B., & Ma, Y. (2022). Application of UAV multisensor data and ensemble approach for high-throughput estimation of maize phenotyping traits. Plant Phenomics. Specs—DJI Mavic 3 Enterprise. Retrieved 15 February 2025 from https://ag.dji.com/mavic-3-m
- Wang, K., & Jin, Y. (2023). Mapping Walnut water Stress with High Resolution Multispectral UAV Imagery and Machine Learning. arXiv preprint arXiv:2401.01375.
- Zhao, S., Kang, F., Li, J., & Ma, C. (2021). Structural health monitoring and inspection of dams based on UAV photogrammetry with image 3D reconstruction. Automation in Construction, 130, 103832.
- Zheng, Z., Yuan, J., Yao, W., Kwan, P., Yao, H., Liu, Q., & Guo, L. (2024). Fusion of UAV-Acquired Visible Images and Multispectral Data by Applying Machine-Learning Methods in Crop Classification. Agronomy, 14(11), 2670. https://doi.org/10.3390/agronomy14112670.
There are 32 citations in total.
Details
| Primary Language | English |
|---|---|
| Subjects | Photogrammetry and Remote Sensing |
| Journal Section | Research Articles |
| Authors | |
| Publication Date | March 31, 2025 |
| Submission Date | February 16, 2025 |
| Acceptance Date | February 21, 2025 |
| Published in Issue | Year 2025 Volume: 12 Issue: 1 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.