Swim Sentinel

Swim Senitel is a next-generation water safety early detection system designed to address the silent threat of drowning by combining wearable IoT technology with overhead Computer Vision. To overcome the limitations of human visual monitoring and the high costs of underwater cameras, our system utilizes a hybrid sensor fusion approach. A wearable wristband serves as a personal sensing layer to monitor for abnormal wrist motion and inactivity duration. Simultaneously, a fixed above-water camera continuously monitors the pool to detect distress-like visual patterns and identify the exact pool zone where the incident is occurring. In this architecture, the wearable indicates who may be at risk, while the camera provides clear visual evidence of where they are and what is happening. The system's technical implementation utilizes the YOLO11s model for lightweight object detection and ByteTrack to maintain an anonymous track ID for the individual across multiple frames. To prevent false alarms caused by normal swimming or splashing, a Temporal State Engine evaluates the centroid displacement and requires the distress-like appearance to persist over time before escalating the state to suspected distress or inactivity. This structured visual evidence is then sent directly to a fusion engine, which cross-references it with the data from the wearable device. By fusing these signals, the system outputs actionable data to a dashboard and buzzer, acting as a highly scalable, real-time decision-support tool for lifeguards rather than an independent medical diagnostic device