CALCULATING WAVE OVERTOPPING VOLUMES THROUGH IMAGE CLUSTERING ANALYSIS

In the last decade, the use of video imagery has been increasingly recognised as an efficient technique to model the wave motion (e.g. Buscombe et al., 2020; Stringari et al., 2021) and the wave-structure interaction processes (i.a., Hofland et al., 2015, Den Bieman et al., 2020), overcoming some of the limitations imposed by physical measurements. In the stream of machine-learning techniques for reconstructing quantitative data from video imagery, Formentin et al. (2021) developed a technique based on image clustering analysis to reconstruct the free-surface elevation and the wave overtopping flow characteristics above the crest of coastal dikes. Such methodology was recently extended (Formentin et al., 2023) to estimate also the wave runup at crown walls. Building on this methodology, this paper aims at proposing the videography, and specifically the image-clustering, as effective and accurate technique to estimate also the instantaneous and total wave overtopping volumes.