Integrating multi-source data for canopy gap detection and distribution modeling in a mixed forest ecosystem

Research on the segmentation and analysis of forest canopy gaps (CGs) is important for understanding forest dynamics, biodiversity, and ecosystem resilience. In this study, seasonal changes in CGs over a five-year period were analyzed using unmanned aerial vehicle (UAV) RGB imagery, satellite multispectral imagery, Synthetic Aperture Radar (SAR), and Light Detection and Ranging (LiDAR) data. The aerial LiDAR achieved the highest segmentation accuracy (87%) for smaller gaps, followed by UAV RGB imagery (84%) and satellite-based observations (70%). Statistical models, including the Weibull Distribution and the Markov Chain, were used for spatial and temporal modeling of CGs. Gap size distribution over the five years showed that smaller CGs (< 100 m²) were more common in the early years, while larger gaps increased in later years, particularly in spring and autumn. This research supports CG segmentation and analysis by combining multi-source data and statistical modeling. Providing a flexible solution for monitoring forest ecosystems and promoting sustainable management practices.