RipAlert Dataset: Annotated Coastal Imagery for Early-Stage Rip Current Detection

This dataset, curated for the RipAlert framework, comprises 1768 annotated frames for object detection. It is specifically curated to capture subtle, early-stage rip current patterns and support predictive modeling.The data was assembled by combining publicly available coastal footage with proprietary video collected via drone and shore surveillance in high-risk areas. Annotations include bounding boxes and confidence levels for rip current identification. This dataset is publicly released to promote reproducible research and advance AI applications for coastal safety.