We propose a semantic segmentation model to analyze coast- al dynamics. Using advanced AI techniques, precise segmenta- tion masks are generated, overcoming challenges like changing weather conditions, glare, or shadows. A diverse dataset ensures adaptability, classifying features such as waves, sand, foam, and static infrastructures at the pixel level. This enables detailed anal- ysis of interactions between marine elements and coastal struc- tures, and can lead to measurements such as wave period, crucial for predicting overtopping events and identifying abnormal sea behavior. Experiments at Las Canteras Beach in Gran Canaria, a location where our model was not trained, yet it still performed well, demonstrate its effectiveness. This research illustrates AI's potential in advancing coastal management and environmental monitoring.
Reyes et al. (Wed,) studied this question.