Machine Learning Approach for LiDAR-based Tree Species Classification in Forest Ecosystem Mapping
Keywords:
Keywords: Tree species classification, Forest mapping, Machine Learning, LiDAR Data, Logistic Regression.Abstract
Accurate tree species classification in forest ecosystems is crucial for biodiversity conservation, forestmanagement, and ecological research. Traditionally, forests have been mapped through labour-intensivefield surveys and the visual interpretation of aerial images, methods prone to human error andinefficiencies. Studies show that error rates in manual mapping can exceed 15%, with variability inexpertise and limited scalability contributing to inconsistencies
References
Colgan, M. S., Baldeck, C. A., Féret, J. B., and Asner, G. P., "Mapping savanna tree species at ecosystem scales using support vector machine classification and BRDF correction on airborne hyperspectral and LiDAR data," Remote Sensing, 3462-3480.
George, R., Padalia, H., and Kushwaha, S. P. S., "Forest tree species discrimination in western Himalaya using EO-cedar cypresscedar 6897989.7%cypress 86 68288.8%precision88.9% 89.6% avg. f-value 89.3%classification resultrecallactual85.3 91.4 89.5 91.6 85.9 80.085.090.095.00 3 3 6 6 9 9 12 12 15accuracy(%)distance(m)Classification accuracy - distance," International Journal of Applied Earth observation and Geoinformation,
-149.
