Edge-side multiparameter non-contact health monitoring system based on rPPG and infrared sensor

Non-contact vital signs monitoring is of great importance in medical care and public health. However, existing methods often rely on single sensors or high-power platforms, making it difficult to balance real-time performance, energy efficiency, and robustness. This paper presents an edge-side monitoring system based on video remote photoplethysmography (rPPG), integrated with a K210 neural network processor and an STM32 microcontroller to achieve real-time estimation of heart rate, with MAX30102 used for reference validation. Experimental results show that the system achieves high accuracy in heart rate measurement, with relative errors within acceptable ranges, and operates stably on a low-power platform. The work applies lightweight optimization in hardware and algorithms, providing a reproducible engineering solution for low-power, real-time, non-contact health monitoring in community and elderly care scenarios.