Alternative thermal calibrations of focal plane array infrared cameras

Abstract

In the past decades, the FPA technology has completely and definitively supplanted the scanning infrared cameras; however, if these devices allow better spatial resolution, they induce several calibration problems due to the non-uniformity of the detectors array. For several applications considering the temperature field and not only a mean value (e.g. to realise maps of dissipative sources in thermomechanical engineering, to determine local thermophysical properties…), one can not only consider the global calibration proposed by the manufacturers, and some home-made specific procedure must be developed. Moreover, these built-in calibration are often associated to Non-uniformity correction (NUC) and Bad pixel replacement procedures, that induce a local correlation between pixels, that is incompatible with any data treatment including spatial derivatives (gradient, Laplacian); then, pixel to pixel calibration is the only solution for these kind of applications. In order to help the users needing such a measurement quality to choose how to calibrate their Focal Plane Array camera, different methods for calibration are proposed and tested in this paper. A comparison between these different calibrations and the standard built-in calibration including a non-uniformity correction (NUC) has been realized, and it has been shown that, in each configuration, a pixel to pixel calibration gives better results that global one.