Accurate classification of wetland vegetation is essential for biodiversity conservation and carbon cycle monitoring, yet traditional methods have struggled with the complex vegetation composition and similar canopy spectra among species in karst wetlands. Researchers from Guilin University of Technology and collaborators have developed a solution that significantly advances this field, publishing their findings in Journal of Remote Sensing on October 16, 2025.
The study introduces an adaptive ensemble learning stacking (AEL-Stacking) framework that merges hyperspectral imagery and LiDAR point-cloud data collected by unmanned aerial vehicles. This approach achieved up to 92.77% accuracy in vegetation species identification, substantially outperforming traditional models by up to 9.5%. The research demonstrates how spectral and structural features jointly improve ecosystem mapping and restoration strategies for these globally significant ecosystems that regulate water, store carbon, and harbor rich biodiversity.
Field surveys were conducted in the Huixian Karst Wetland of Guilin, China, where UAV flights equipped with specialized sensors collected over 4,500 hyperspectral images and dense point clouds covering 13 vegetation types. The integrated dataset included species such as lotus, miscanthus, and camphor trees. Through recursive feature elimination and correlation analysis, researchers selected 40 optimal features from more than 600 variables, with LiDAR-derived digital surface model variables proving particularly important for distinguishing species with distinct vertical structures.
The AEL-Stacking model integrates Random Forest, LightGBM, and CatBoost classifiers under a grid-search-optimized adaptive framework, using 70% of data for training and 30% for testing with 10-fold cross-validation. The framework also employs local interpretable model-agnostic explanations to visualize how each feature contributes to decision-making, offering both high precision and interpretability in mapping complex wetland vegetation structures. According to the study published at https://spj.science.org/doi/10.34133/remotesensing.0452, this approach significantly reduced misclassification between morphologically similar species, with lotus and miscanthus achieving classification F1-scores above 0.9.
"Our approach bridges the gap between spectral and structural sensing," said Dr. Bolin Fu, corresponding author of the study. "By combining UAV hyperspectral and LiDAR data through adaptive ensemble learning, we achieved both precision and interpretability in vegetation mapping. The framework not only improves species recognition in complex karst environments but also provides a generalizable tool for ecological monitoring and habitat restoration worldwide."
The research highlights the synergy between optical and structural data in resolving species with overlapping spectral signatures. Hyperspectral vegetation indices such as NDVI and blue-edge parameters enhanced recognition of herbaceous species, while LiDAR features were pivotal for distinguishing species with distinct vertical structures. This integrative framework demonstrates a scalable and explainable approach for high-resolution wetland mapping that could potentially be applied to forest, grassland, and coastal ecosystems.
Future work will focus on integrating multi-temporal UAV observations and satellite data fusion to monitor seasonal vegetation dynamics and climate-driven changes in wetland health. By enhancing the transparency and accuracy of AI-driven ecological models, this research supports the global agenda for biodiversity conservation and carbon neutrality while providing detailed vegetation maps critical for ecosystem monitoring. The study was supported by multiple funding sources including the National Natural Science Foundation of China and the Natural Science Foundation of Guangxi Zhuang Autonomous Region.
