A Real-Time Human Position Estimation System Using a Low-Resolution Infrared Thermal Sensor Array

Human detection and position estimation technologies are widely applied in various fields, such as monitoring systems for residential safety or real-time detection of patient movements in healthcare facilities such as nursing homes to ensure their safety. Current monitoring systems are primarily based on vision-based technologies, such as traditional cameras. However, these systems often fail in low-light conditions or under other external influences, resulting in low image clarity and hindering subsequent processing. Alternatively, specific sensors can address these issues. An infrared thermal array sensor is employed in this study. Although this sensor lacks the resolution to identify individual identities, it still provides essential data for accurate location estimation, thereby preserving personal privacy. In this research, a human position detection system based on a low-resolution infrared thermal sensor is proposed. The system uses an AMG8833 thermal sensor connected to a Raspberry Pi, which captures and processes data in real time and performs computations directly on the Raspberry Pi. Our system effectively detects and estimates the positions of multiple human targets, demonstrating excellent performance even in scenarios with continuous human movement.