Individual Tree Crown Detection and Classification of Live and Dead Trees Using a Mask Region-Based Convolutional Neural Network (Mask R-CNN)

Forests, Journal Year: 2024, Volume and Issue: 15(11), P. 1900 - 1900

Published: Oct. 28, 2024

Mapping the distribution of living and dead trees in forests, particularly ecologically fragile areas where forests serve as crucial ecological environments, is essential for assessing forest health, carbon storage capacity, biodiversity. Convolutional neural networks, including Mask R-CNN, can assist rapid accurate monitoring. In this study, R-CNN was employed to detect crowns Casuarina equisetifolia distinguish between live Pingtan Comprehensive Pilot Zone, Fujian, China. High-resolution images five plots were obtained using a multispectral Unmanned Aerial Vehicle. Six band combinations derivatives, RGB, RGB-digital surface model (DSM), Multispectral, Multispectral-DSM, Vegetation Index, Vegetation-Index-DSM, used tree crown detection classification trees. Five-fold cross-validation divide manually annotated dataset 21,800 7157 into training validation sets, which validating models. The results demonstrate that RGB combination achieved most effective performance (average F1 score = 74.75%, IoU 70.85%). RGB–DSM exhibited highest accuracy 71.16%, 68.28%). lower than trees, may be due similar spectral features across similarity background, resulting false identification. For simultaneous produced promising 74.18%, 69.8%). It demonstrates achieve Our study could provide managers with detailed information on condition, has potential improve management.

Language: Английский

---

The segmentation of debris-flow fans based on local features and spatial attention mechanism

Journal of Geographical Sciences, Journal Year: 2024, Volume and Issue: 34(12), P. 2534 - 2550

Published: Dec. 1, 2024

Language: Английский

---