Non-Contact Body Temperature Monitoring Using Low-Cost Thermal Cameras in Smart Home Environments: An Experimental Study and Prototype Implementation

This paper investigates the feasibility of using low-cost thermal cameras for continuous, non-contact body temperature monitoring within a smart home environment. Three thermal imaging devices were evaluated: two MLX90640 sensor modules mounted on different breakout boards (Pimoroni MLX-P and RobotShop MLX-R) and the RS PRO T-10 (RS-T10). Experiments were conducted in a controlled smart home laboratory with 30 participants across distances of 1 to 5 metres and ambient temperatures ranging from 16°C to 25°C. Physical factors, including skin colour emissivity, eyeglass use, and illumination source, were systematically assessed. Results demonstrate that ambient temperature and measurement distance are the dominant sources of measurement error, whereas skin emissivity and illumination type have a negligible impact. A bilinear interpolation compensation model was developed and integrated into a Three-Level Smart Home Architecture (3-LSHA), producing estimated body temperature values with an acceptable deviation from baseline readings across all tested conditions. A prototype system was implemented to demonstrate end-to-end well-being monitoring across multiple rooms. The findings confirm that low-cost thermal cameras, when coupled with appropriate compensation algorithms, are viable sensors for ambient-assisted living applications targeting body temperature trend monitoring and health alerting. Keywords: Ambient Assisted Living, Bilinear Interpolation, Body Temperature Monitoring, llumination Infrared Thermography, MLX90640, Smart Home, Thermal Camera, Well-Being Monitoring.