Area 2 — AI-Driven Tree Analysis and Simulation

Large-scale data collection only creates value if it can be efficiently analyzed. Our lab develops deep learning models that extract ecologically meaningful information from 3D point cloud data — detecting individual trees, estimating their dimensions and biomass, classifying species, and simulating tree growth trajectories over time.

Recent work includes AI models for reconstructing the three-dimensional architecture of individual trees, mapping urban tree cover at national scales, and building super-resolution algorithms that enhance low-density LiDAR data and coarse satellite imagery for individual-tree-level analysis. The lab also developed iForester, a smartphone application that allows landowners, students, and the public to measure trees and estimate timber value using only a phone — making digital forestry accessible far beyond the research community.

This image represents multiple computer alterations of a tree landscape.

Publication

  1. Shao, J., Choi, D.H., Liu, J., Tian, X., Thapa, B., Lee, S., Habib, A., & Fei, S. (2026). A three-stage framework for stand-level automated stem volume estimation in temperate forests using mobile laser scanning. Remote Sensing of Environment, 335, 115246. https://doi.org/10.1016/j.rse.2026.115246
  2. Jung, M., Choi, J., Carpenter, J., Fei, S., & Jung, J. (2025). Individual tree biomass estimation using single-scan terrestrial laser scanner with efficient projection-based deep learning. Journal of Forestry. https://doi.org/10.1007/s44392-025-00065-6
  3. Zhou, X., Li, B., Benes, B., Habib, A., Fei, S., Shao, J., & Pirk, S. (2025). TreeStructor: Forest reconstruction with neural ranking. IEEE Transactions on Geoscience and Remote Sensing, 63, 4408419. https://doi.org/10.1109/TGRS.2025.3558312
  4. Carpenter, J., Jung, M., Goel, A., Fei, S., & Jung, J. (2025). Species classification of northern hardwood forest inventories from terrestrial laser scans and airborne LiDAR. Frontiers in Forests and Global Change, 8, 1500178. https://doi.org/10.3389/ffgc.2025.1500178
  5. Carpenter, J., Jung, J., Oh, S., Hardiman, B., & Fei, S. (2022). An unsupervised canopy-to-root pathing (UCRP) tree segmentation algorithm for automatic forest mapping. Remote Sensing, 14(17), 4274. https://doi.org/10.3390/rs14174274