Visual ocean wave analytics

Abstract

Accurate prediction and tracking of ocean wave dynamics especially near the beach side are important for enhancing the safety and experience of surfers. This study presents a deep learning based approach that uses computer vision techniques to forecast wave break timings and positions. The proposed method employs a CNN architecture trained on manually annotated video data to analyze and predict wave break and positions. The model is designed to provide realistic predictions within a short time frame of few seconds, enabling surfers to assess the conditions and make decisions about whether to paddle or not. This study also highlights the potential of integrating deep learning into ocean wave analytics for real-time applications.