A remote sensing image object detection algorithm based on transfer learning

Abstract

Addressing the limitations of conventional heart rate detectors, which solely monitor heart rate without providing alerts during critical incidents, a multifunctional heart rate detection system has been developed. This system incorporates a GD32 as the main controller and utilizes a pulse sensor for heart rate monitoring, an MLX90614 infrared temperature sensor for body temperature measurements, an MPU6050 acceleration sensor for motion posture detection, and a SIM868 module for GPS positioning and server communication. This comprehensive setup allows for real-time monitoring of the wearer's heart rate and body temperature, assessment of fall incidents based on motion posture data, and regular uploading of health data to a server. Additionally, the system is equipped with a one-touch emergency call feature. In the event of abnormal heart rate readings or a fall, it triggers an alarm, acquires GPS coordinates, and automatically dials an emergency contact number. After thorough testing and verification, the system has demonstrated full functionality with heart rate data accuracy within the expected error range, offering a reliable safety mechanism for patients while on the move.