Proposal and verification of rip current detection using AI

Conference paper (2021)

Authors

T. Chuo University, Japan
R. Chuo University, Japan
R. Chuo University, Japan
T. Chuo University, Japan
Copyright: © the authors
Coastal observations and monitoring Coastal management and risk assessments
More Info
expand_more

Published Date

28-06-2021

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

There are from 2,000 to 3,000 rescues including those of unconscious people every year on the beaches of Japan, as shown in Fig. 1. The occurrence of drowning accidents is mainly caused by the rip current (Ishikawa et al., 2014), it accounts for 48 % of drowning accidents, as shown in Fig. 2. Also, in Australia, the United States and the United Kingdom, more than 50 % of rescue accidents are caused by rip currents (Brighton et al., 2013). In order to reduce the rip current accidents, beach users need to recognize rip currents, then they have to avoid them using risk assessment. However, it is the difficulty of risk recognition and judgement under the momentary change in natural phenomenon for beach users. Especially, almost all beach users understand the risk in the case of high wave conditions due to easy visual understanding, whereas they cannot understand rip currents the same way. On the other hand, swimming areas along the shore are very limited, however the number of lifeguards is small at around 1 lifeguard compared to the thousands of beach users. In addition, beach users sometimes enter unpatrolled areas outside the swimming areas. Therefore, we developed a new technology that can automatically detect the rip currents by the Artificial Intelligence (AI), and notify beach users and lifeguards using the Internet of Things (IoT). In this study, we verified the accuracy of the rip current detection by the AI, using a field measurement, an image analysis and a numerical simulation. Also, we examined the log data of 2019 that was actually operated.