In this paper, application of uncooled thermal imaging for human body temperature screening is evaluated in terms of accuracy, measurement uncertainty and reliability. Physical theory of contactless thermometry is first analysed for generalisations and assumptions, which often deviate from actual conditions and are practically impossible to assure. Human is further analysed as a measured subject, identifying eye cantus as the most appropriate facial measurement spot. Accuracy of skin temperature measurement and core body temperature assessment is calculated, taking into account both the measurement instrument and measured subject, the human.
Results of measurement uncertainty analysis, gathered to the extent of available data, are added to common measurement uncertainty of the core body temperature measurement and assessment process. Theoretical combined uncertainty of the measurement is considerably higher than uncertainties claimed by manufacturers and required by medical specialists at ±0,3 °C and can be as high as ±2 °C.