Monitoring early-successional trees for tropical forest restoration using low-cost UAV-based species classification

Logged forests cover four million square kilometers of the tropics, capturing carbon more rapidly than temperate and harboring rich biodiversity. Restoring these is essential to help avoid worst impacts climate change. Yet monitoring tropical forest recovery challenging. We track abundance early-successional species in a restoration concession Indonesia. If are carefully chosen, they can be used as an indicator progress. present SLIC-UAV, new pipeline for processing Unoccupied Aerial Vehicle (UAV) imagery using simple linear iterative clustering (SLIC)to map forests. The comprises: (a) field verified approach manually labeling species; (b) automatic segmentation into “superpixels” (c) machine learning classification based on both spectral textural features. Creating superpixels massively reduces dataset's dimensionality enables use features, which improve accuracy. In addition, this flexible with regards spatial distribution training data. This allowed us collect high-quality data local experts. accuracy ranged from 74.3% four-species task 91.7% when focusing only key early-succesional species. then extended models across 100 hectares forest, mapping dominance condition entire project.