第137回日本森林学会大会 発表検索
講演詳細
経営部門[Forest Management]
| 日付 | 2026年3月18日 |
|---|---|
| 開始時刻 | ポスター発表 |
| 会場名 | 多目的ホール |
| 講演番号 | PD-20 |
| 発表題目 | Integrating UAV Multispectral and LiDAR Data for Forest Carbon Stock Estimation in Silvicultural Treatment Areas Integrating UAV Multispectral and LiDAR Data for Forest Carbon Stock Estimation in Silvicultural Treatment Areas |
| 所属 | The University of Tokyo |
| 要旨本文 | Accurate large-scale estimation of living biomass carbon stocks remains challenging in managed forests with limited field data. We developed plot-trained machine learning models integrating UAV multispectral imagery with UAV- and airborne LiDAR structural metrics to estimate carbon stocks across silvicultural treatment areas in two study sites in Hokkaido, northern Japan: a typhoon-disturbed forest and natural mixed forests. Field data from 38 plots were subdivided into 10 × 10 m grid cells for consistent feature extraction and wall-to-wall prediction. Spectral and structural features were extracted, and plot-level carbon stocks were allocated to grid cells and re-aggregated for evaluation. Random Forest (RF) and Extreme Gradient Boosting (XGBoost) were evaluated using leave-one-plot-out cross-validation. The results indicate that integrated UAV multispectral and LiDAR data enable scalable carbon stock estimation, with XGBoost outperforming RF (RMSE 27 vs. 34 Mg C ha__). |
| 著者氏名 | ○Nyo Me Htun1 ・ Toshiaki Owari2 ・ Satoshi N. Suzuki3 ・ Songqiu Deng2 ・ Tetsuyuki Kobayashi2 ・ Sakura Asato2 ・ Akio Oshima2 ・ Mutsuki Hirama2 ・ Koichi Takahashi2 ・ Yasuo Isozaki2 ・ Takumi Okahira2 ・ Ryota Konda4 ・ Satoshi Kita5 ・ Manato Fushimi4 |
| 著者所属 | 1The University of Tokyo, Forest GX/DX Co-creation Center ・ 2The University of Tokyo Hokkaido Experimental Forest ・ 3Nakagawa Experimental Forest, Field Science Center for Northern Biosphere, Hokkaido University ・ 4Tsukuba Research Institute, Sumitomo Forestry Co., Ltd. ・ 5Forest and Landscape Research Center, Sumitomo Forestry Co., Ltd. |
| キーワード | UAV multispectral imagery, LiDAR-derived DTM, carbon stock, silvicultural treatment areas |
| Key word | UAV multispectral imagery, LiDAR-derived DTM, carbon stock, silvicultural treatment areas |