A Machine Learning Approach for Predicting Canopy Base Height from Airborne LiDAR

Yıl 2025, Cilt: 21 Sayı: 2, 304 - 316, 30.12.2025

Tufan Demirel

https://doi.org/10.58816/duzceod.1801720

Öz

Forest ecosystems are critically important for carbon sequestration, biodiversity, and other ecosystem services. In this context, accurate and up-to-date measurements of forest structural parameters are essential for sustainable forestry and ecosystem management. Canopy Base Height (CBH) is a key parameter for characterizing the vertical structure of forests and plays a significant role in biomass and log quality models as well as forest fire simulation models. However, conventional field measurements are time-consuming and often impractical in areas with difficult access. In this study, a CBH prediction model was developed for pure maritime pine (Pinus pinaster) stands using airborne LiDAR data and the XGBoost machine learning algorithm. Field measurements were conducted in 32 sample plots, and CBH values were calculated for these plots. 56 elevation and point cloud metrics were extracted from aerial LiDAR data, and highly correlated variables were removed. The remaining 23 metrics were used for modeling. Hyperparameter optimization for the XGBoost model was performed using Grid Search, with the best performance achieved using nrounds=50, max_depth=3, and eta=0.3. The model’s generalization performance was evaluated via LOOCV, demonstrating high accuracy (R²=0.80; RMSE=0.28 m; MAE=0.21 m), and predictions showed strong agreement with observed CBH values. Variable importance analysis revealed that the elev_skewness, elev_max, and elev_AIH_30th metrics had significant effects on CBH. These results indicate that the XGBoost algorithm is a reliable and effective method for predicting CBH based on LiDAR data.

Anahtar Kelimeler

XGBoost , maritime pine , Remote Sensing

Hava Tabanlı LiDAR ile Tepe Altı Yüksekliğinin Tahmininde Makine Öğrenmesi Tabanlı Bir Yaklaşım

Öz

Orman ekosistemleri, karbon tutma, biyolojik çeşitlilik ve diğer ekosistem hizmetleri açısından kritik öneme sahiptir. Bu bağlamda, orman yapısal parametrelerinin doğru ve güncel ölçümü, sürdürülebilir ormancılık ve ekosistem yönetimi için gereklidir. Tepe altı yüksekliği (TAY), ormanların dikey yapısını tanımlamada temel bir parametre olup, biyokütle ve tomruk kalitesi modelleri ile orman yangını simülasyon modellerinde önemli bir rol oynar. Ancak klasik arazi ölçümleri zaman alıcı ve erişimi güç alanlarda uygulanabilirlik açısından sınırlıdır. Bu çalışmada, saf sahil çamı (Pinus pinaster) meşcerelerinde Hava tabanlı LiDAR verileri ve makine öğrenmesi algoritmalarından XGBoost kullanılarak TAY tahmin modeli geliştirilmiştir. 32 örnek alanda arazi ölçümleri gerçekleştirilmiş ve örnek alanlara ait TAY değerleri hesaplanmıştır. Hava tabanlı LiDAR verilerinden 56 yükseklik ve nokta bulutu metrikleri türetilmiş, yüksek korelasyonlu değişkenler elenerek 23 metrik modelde kullanılmıştır. XGBoost modeli için hiperparametre optimizasyonu Grid Search yöntemiyle yapılmış ve en iyi performans nrounds=50, max_depth=3 ve eta=0,3 parametreleri ile elde edilmiştir. Modelin genelleme performansı LOOCV ile değerlendirildiğinde yüksek doğruluk sağlanmış (R²=0,80; RMSE=0,28 m; MAE=0,21 m) ve tahminler gözlenen TAY değerleri ile güçlü uyum göstermiştir. Değişken önem analizi, elev_skewness, elev_max ve elev_AIH_30th metriklerinin TAY üzerinde önemli bir etkiye sahip olduğunu ortaya koymuştur. Sonuçlar, XGBoost algoritmasının LiDAR verilerine dayalı TAY tahmininde güvenilir ve etkin bir yöntem olduğunu göstermektedir.

Anahtar Kelimeler

XGBoost , Sahil çamı , Uzaktan algılama

Kaynakça

• Andersen, H., McGaughey, R. J. ve Reutebuch, S. E. (2005). Estimating forest canopy fuel parameters using LIDAR data. Remote Sensing of Environment, 94(4), 441–449. https://doi.org/10.1016/j.rse.2004.10.013
• Bahadir, M., Karsli, F., Yildirim, F. S. ve Misir, M. (2025). Tree crown segmentation and estimation of metrics from point clouds with improved local maximum method. The Photogrammetric Record, 40(191). https://doi.org/10.1111/phor.70015
• Botequim, B., Fernandes, P. M., Borges, J. G., González-Ferreiro, E. ve Guerra-Hernández, J. (2019). Improving silvicultural practices for Mediterranean forests through fire behaviour modelling using LiDAR-derived canopy fuel characteristics. International Journal of Wildland Fire, 28(11), 823. https://doi.org/10.1071/wf19001
• Chen, T. ve Guestrin, C. (2016). XGBoost: A scalable tree boosting system. KDD '16: proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining içinde (ss. 785 – 794). https://doi.org/10.1145/2939672.2939785
• Coskuner, K. A., Vatandaslar, C., Ozturk, M., Harman, I., Bilgili, E., Karahalil, U., Berber, T. ve Gormus, E. T. (2023). Estimating Mediterranean stand fuel characteristics using handheld mobile laser scanning technology. International Journal of Wildland Fire, 32(9), 1347–1363. https://doi.org/10.1071/wf23005
• Dalla Corte, A. P. D., Souza, D. V., Rex, F. E., Sanquetta, C. R., Mohan, M., Silva, C. A., Zambrano, A. M. A., Prata, G., De Almeida, D. R. A., Trautenmüller, J. W., Klauberg, C., De Moraes, A., Sanquetta, M. N., Wilkinson, B. ve Broadbent, E. N. (2020). Forest inventory with high-density UAV-Lidar: Machine learning approaches for predicting individual tree attributes. Computers and Electronics in Agriculture, 179, 105815. https://doi.org/10.1016/j.compag.2020.105815
• Dean, T. J., Cao, Q. V., Roberts, S. D. ve Evans, D. L. (2009). Measuring heights to crown base and crown median with LiDAR in a mature, even-aged loblolly pine stand. Forest Ecology and Management, 257(1), 126-133. https://doi.org/10.1016/j.foreco.2008.08.024
• Erdody, T. L. ve Moskal, L. M. (2010). Fusion of LiDAR and imagery for estimating forest canopy fuels. Remote Sensing of Environment, 114(4), 725–737. https://doi.org/10.1016/j.rse.2009.11.002
• FAO. (2023). The State of the World’s Forests 2023: Forests and health. Food and Agriculture Organization of the United Nations.
• García, M., Saatchi, S., Ustin, S. ve Balzter, H. (2018). Modelling forest canopy height by integrating airborne LiDAR samples with satellite Radar and multispectral imagery. International Journal of Applied Earth Observation and Geoinformation, 66, 159–173. https://doi.org/10.1016/j.jag.2017.11.017
• Green Valley International. (2023). LiDAR360 user guide. GreenValley International Ltd. https://www.greenvalleyintl.com/gvi/web/us/file/EN-User-Guide-LiDAR360.pdf
• Guo, Q., Su, Y., Hu, T., Guan, H., Jin, S., Zhang, J., Zhao, X., Xu, K., Wei, D., Kelly, M. ve Coops, N. C. (2021). LIDAR boosts 3D Ecological Observations and Modelings: A Review and perspective. IEEE Geoscience and Remote Sensing Magazine, 9(1), 232–257. https://doi.org/10.1109/mgrs.2020.3032713
• Hall, S. A., Burke, I., Box, D. O., Kaufmann, M. R. ve Stoker, J. M. (2005). Estimating stand structure using discrete-return lidar: an example from low density, fire prone ponderosa pine forests. Forest Ecology and Management, 208(1–3), 189–209. https://doi.org/10.1016/j.foreco.2004.12.001
• Hermosilla, T., Ruiz, L. A., Kazakova, A. N., Coops, N. C. ve Moskal, L. M. (2014). Estimation of forest structure and canopy fuel parameters from small-footprint full-waveform LiDAR data. International Journal of Wildland Fire, 23(2), 224. https://doi.org/10.1071/wf13086
• Kaushal, S. ve Baishya, R. (2021). Stand structure and species diversity regulate biomass carbon stock under major Central Himalayan forest types of India. Ecological Processes, 10(1). https://doi.org/10.1186/s13717-021-00283-8
• Kelly, M., Su, Y., Di Tommaso, S., Fry, D., Collins, B., Stephens, S. ve Guo, Q. (2017). Impact of error in LIDAR-derived canopy height and canopy base height on modeled wildfire behavior in the Sierra Nevada, California, USA. Remote Sensing, 10(1), 10. https://doi.org/10.3390/rs10010010
• Koukoulas, S. ve Blackburn, G. A. (2005). Spatial relationships between tree species and gap characteristics in broad‐leaved deciduous woodland. Journal of Vegetation Science, 16(5), 587-596. https://doi.org/10.1111/j.1654-1103.2005.tb02400.x
• Kucuk, O., Goltas, M., Demirel, T., Mitsopoulos, I. ve Bilgili, E. (2021). Predicting canopy fuel characteristics in Pinus brutia Ten., Pinus nigra Arnold and Pinus pinaster Ait. forests from stand variables in north-western Turkey. Environmental Engineering and Management Journal, 20(2), 309–318. https://doi.org/10.30638/eemj.2021.031
• Küçük, Ö., Bilgili, E., Durmaz, B. D., Sağlam, B. ve Baysal, İ. (2009). Örtü yangınının tepe yangınına geçişinde etkili olan faktörler. Kastamonu University Journal of Forestry Faculty, 9(2), 80-85.
• Lee, H., Slatton, K. C., Roth, B. E. ve Cropper, W. P. (2010). Adaptive clustering of airborne LiDAR data to segment individual tree crowns in managed pine forests. International Journal of Remote Sensing, 31(1), 117–139. https://doi.org/10.1080/01431160902882561
• Luo, L., Zhai, Q., Su, Y., Ma, Q., Kelly, M. ve Guo, Q. (2018). Simple method for direct crown base height estimation of individual conifer trees using airborne LiDAR data. Optics Express, 26(10), A562. https://doi.org/10.1364/oe.26.00a562
• Maguya, A., Tegel, K., Junttila, V., Kauranne, T., Korhonen, M., Burns, J., Leppanen, V. ve Sanz, B. (2015). Moving voxel method for estimating canopy base height from airborne laser scanner data. Remote Sensing, 7(7), 8950–8972. https://doi.org/10.3390/rs70708950
• Maltamo, M., Karjalainen, T., Repola, J. ve Vauhkonen, J. (2018). Incorporating tree- and stand-level information on crown base height into multivariate forest management inventories based on airborne laser scanning. Silva Fennica, 52(3). https://doi.org/10.14214/sf.10006
• Næsset, E. ve Økland, T. (2002). Estimating tree height and tree crown properties using airborne scanning laser in a boreal nature reserve. Remote Sensing of Environment, 79(1), 105–115. https://doi.org/10.1016/s0034-4257(01)00243-7
• Stefanidou, A., Gitas, I., Korhonen, L., Stavrakoudis, D. ve Georgopoulos, N. (2020). LIDAR-Based estimates of canopy base height for a dense uneven-aged structured forest. Remote Sensing, 12(10), 1565. https://doi.org/10.3390/rs12101565
• Valentine, H. T., Amateis, R. L., Gove, J. H. ve Mäkelä, A. (2013). Crown-rise and crown-length dynamics: application to loblolly pine. Forestry, 86(3), 371-375. https://doi.org/10.1093/forestry/cpt007
• Vauhkonen, J. (2010). Estimating crown base height for Scots pine by means of the 3D geometry of airborne laser scanning data. International Journal of Remote Sensing, 31(5), 1213-1226. https://doi.org/10.1080/01431160903380615
• Viedma, O., Silva, C. A., Moreno, J. M. ve Hudak, A. T. (2024). LadderFuelsR: A new automated tool for vertical fuel continuity analysis and crown base height detection using light detection and ranging. Methods in Ecology and Evolution 15(11), 1958-1967. https://doi.org/10.1111/2041-210x.14427
• Zhao, K., Popescu, S., Meng, X., Pang, Y. ve Agca, M. (2011). Characterizing forest canopy structure with lidar composite metrics and machine learning. Remote Sensing of Environment, 115(8), 1978-1996. https://doi.org/10.1016/j.rse.2011.04.001