Spectroradiometric calibration of blackbody sources

摘要

An IR camera responds to infrared radiant energy over a waveband determined by the camera optics and detector. Most cameras operate in either the 3 to 5 mu m or 8 to 12 mu m wavebands as they represent good atmospheric windows. Temperature measurement using these cameras is performed within the camera, which will correct for target emissivity and background temperatures. The algorithm that does this makes a key assumption: the target is a graybody source (constant emissivity over the waveband of the IR camera). To effect this calculation, modern infrared (IR) cameras are calibrated using blackbody sources. The calibration data set is stored in the camera firmware as a lookup table. To be accurate, blackbody sources must be graybodies with emissivities very close to one. The fact that there are no perfect blackbodies can be overcome as long as the emissivity is constant with wavelength. Spectroradiometric calibration of blackbody sources is the best way to ensure the radiant energy emanating from the source follows Planck's law over the waveband of the IR camera being calibrated (another way of saying graybody). At FLIR Systems, Inc. we used a CI Systems SR-5000 spectroradiometer to spectrally characterize 20 blackbody sources of various manufacturer and type. The spectral data were plotted compared to the Planck curve at the same temperature to see how "gray" the blackbody sources were. We also calculated an equivalent temperature that an IR camera would "see" due to any non-gray source behavior, by mathematically answering the question: What temperature would a blackbody need to be to produce the equivalent amount of radiant energy over a given waveband?The paper shows the experimental and mathematical approach to this problem. Results show cavity and "hybrid cavity" blackbodies perform best, with flat plate blackbodies performing well in most cases. The higher temperature flat plate blackbodies had some significant deviations from grayness in some wavebands.