Mapping the Underwater Forest: A Deep Learning Approach to Seagrass Mapping Distribution in Calatagan, Batangas, Philippines, Using Sentinel-2 Satellite Imagery

Abstract

Around 71 percent of the earth’s surface is covered by water primarily saltwater found in the oceans which are essential to the survival of a variety of marine ecosystems. Particularly considering the environmental issues facing the Philippines this study highlights the significance of seagrass beds essential but usually disregarded ecosystems that support the maintenance of water quality and carbon sequestration. To produce an accurate map of the distribution of seagrass in Calatagan, Batangas, high-resolution sentinel-2 imagery is analyzed using deep learning and advanced remote sensing techniques. The approach which includes spectral band selection, data collection and model training produces a deep learning model with an F1 score of 85 percent, precision of 86 percent and overall accuracy of 97.31 percent. The efficiency of remote sensing in monitoring vital coastal habitats in the face of increasing human threats is demonstrated by this study. Through the combination of deep learning algorithms and remote sensing technology this work offers a novel approach to improve ecological analysis in coastal management. This study provides an important step in maintaining and preserving biodiversity by integrating scientific findings into practical conservation plans and strategies.