Speaker
Description
Infrared cameras are widely used in scientific research and industrial applications for temperature measurement and thermal analysis. For quantitative purposes, it is essential to adopt an appropriate radiometric model that accurately represents both the measurement conditions and specific properties of the target object. The diffuse-gray body approximation is commonly used in conventional thermography and is often embedded in camera software. However, this simplification can lead to significant errors when spectral–directional effects are present, such as in temperature monitoring during machining processes, gas quantification, or the thermal analysis of photovoltaic systems. In contrast, the standard formulation to address radiative phenomena is the spectral–directional approach, which captures both angular and spectral dependencies and incorporates critical parameters such as the detector’s spectral response and the atmosphere’s spectral transmissivity. In this work, we simulate the radiance spectra reaching a Long-Wave Infrared (LWIR) camera detector for a given object temperature and emissivity distribution, and analyze the differences between spectral and diffuse-gray models in determining the object temperature through an inverse analysis. We simulated spectra for object temperatures of 300 K, 400 K, 500 K, 1000 K, and 1200 K, considering various emissivity functions randomly generated within the interval [0, 1], including continuously increasing and decreasing profiles. For each simulated scenario, we defined object, reflected, and atmospheric temperatures, along with the atmospheric spectral transmissivity and the detector response curve. Finally, we computed the object temperature using both the diffuse-gray and spectral models, then compared these results against the known temperatures of the original simulated scenarios. The results show that differences between the retrieved temperatures depend strongly on both the spectral behavior of emissivity and the object temperatures. Profiles with increasing and decreasing emissivity variation yielded larger deviations, reaching up to hundreds degrees Celsius. Moreover, the analysis revealed that these discrepancies are amplified at higher object temperatures and low-emissivity values, while for lower temperatures the differences between the models are less pronounced. Readers can benefit from this article, using it as a framework to check the reliability and the measurement deviations introduced by the diffuse-gray approximation across different target temperatures and emissivity distributions.
