Scholarlink

We developed the AI-based system of rip current detection in 2018 to automatically detect rip currents using artificial intelligence and notify beachgoers and lifeguards through the Internet of Things. In this study, we enhanced the system with three complementary modules: drone-based voice alerts, augmented reality (AR) visualization of hazard zones, and help signal identification using AI. These upgrades address two persistent challenges in beach safety: delayed response times and difficulty in locating distressed individuals. The drone module autonomously navigates to rip current locations detected by AI and broadcasts warnings, reducing response times by more than half in rescue simulations. The AR module overlays GPS-linked hazard icons on mobile screens, enabling beachgoers to perceive risks with meter-level spatial accuracy. For help signal identification, we trained a new model using high-resolution inputs and motion-aware frame differencing, improving the detection of raised-arm gestures under realistic wave conditions. Field trials and on-site surveys with lifeguards and beachgoers confirmed the feasibility and positive reception of all three modules, suggesting strong potential to strengthen coastal safety practices.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adding Features to the AI-Based System of Rip Current Detection for Water Safety

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ryo Shimada,
  • Toshinori Ishikawa,
  • Naoya Fujita,
  • Sarina Wada

摘要

We developed the AI-based system of rip current detection in 2018 to automatically detect rip currents using artificial intelligence and notify beachgoers and lifeguards through the Internet of Things. In this study, we enhanced the system with three complementary modules: drone-based voice alerts, augmented reality (AR) visualization of hazard zones, and help signal identification using AI. These upgrades address two persistent challenges in beach safety: delayed response times and difficulty in locating distressed individuals. The drone module autonomously navigates to rip current locations detected by AI and broadcasts warnings, reducing response times by more than half in rescue simulations. The AR module overlays GPS-linked hazard icons on mobile screens, enabling beachgoers to perceive risks with meter-level spatial accuracy. For help signal identification, we trained a new model using high-resolution inputs and motion-aware frame differencing, improving the detection of raised-arm gestures under realistic wave conditions. Field trials and on-site surveys with lifeguards and beachgoers confirmed the feasibility and positive reception of all three modules, suggesting strong potential to strengthen coastal safety practices.