Calibration of a cooled MWIR camera with a narrow-band filter for quantitative thermography

While cooled infrared cameras with narrow-band filters are widely applied, the influence of such filters on measurement accuracy remains poorly investigated. To address this, we propose a calibration methodology for a cooled MWIR camera equipped with a narrow-band filter, including corrections for the filter’s parasitic radiation and a full uncertainty analysis to ensure metrological robustness. The method models thermal radiation using a line-by-line spectral approach and was applied to a filter with a central wavelength of 4,265 ±20 nm and a half-width of 80 ± 15 nm, suitable for InSb detectors and effective for atmospheric CO2 detection. A comparative study was conducted for different integration times and background temperatures to identify the configuration that maximizes sensitivity and reliability. Results show that the proposed compensation for the filter’s parasitic radiation is effective in mitigating systematic errors in radiometric measurements, and that the camera demonstrates the capability to detect a wide range of CO2 concentrations despite filter-induced noise. Furthermore, higher integration times combined with elevated background temperatures produced lower Noise Equivalent Radiance (NER) values and smaller deviations in radiometric results, suggesting that this configuration may eliminate the need for parasitic radiation compensation and thereby reduce calibration uncertainty, which is particularly relevant when working with low concentrations.