Estimating forest degradation using satellite imagery and time series analysis

Keywords: Forest degradation, remote sensing, time series analysis
Abstract Type: Poster Abstract

Abstract

Forest degradation refers to the gradual loss of forest quality and biomass, often resulting from logging, land-use change, and climate impacts, rather than outright deforestation. Satellite imagery enables the detection of subtle changes in forest structure and canopy cover over time, providing a means to track degradation on large scales and in remote areas where ground access may be limited. Time series analysis involves examining sequences of satellite images or other remotely sensed data over time to detect trends and patterns associated with degradation. This approach is particularly valuable because it allows for the differentiation between natural seasonal variations and actual signs of degradation. However, most time series analysis approaches primarily focus on single-pixel information, often ignoring spatial contextual information. In this study, we propose an improved forest degradation estimation model that considers spatial contextual information. Experimental results demonstrate that our model achieves better performance in early degradation detection compared to conventional approaches.