Contactless sleep quality monitoring through thermal vision

Long-term sleep monitoring is essential in many vital health state monitoring applications, including detecting abnormal human behaviour during sleep to support the diagnosis of sleep disorders. Previous approaches to monitoring human behaviour during sleep have relied on inconvenient sensing methods such as wearable devices or intrusive sensing technologies, for example, cameras. This paper explores the potential application of the Thermal Sensor Array (TSA) to monitor human sleep. Humans usually cover their bodies during sleeping, affecting the TSA's ability to sense their infrared radiation. This paper overcomes these challenges by proposing a series of motion-based signal processing rather than processing the human physical appearance in the TSA output. A deterministic approach has been utilised to examine the proposed approach to detect abnormal human movement during sleep.