Page 565 - Sensors and Control Systems in Manufacturing
P. 565

518
                           Cha p te r

                                      Scene Spectral Radiance T e n  x Drifted  x  Original










                                               Spectral Power at Detector
                          FIGURE 10.19  Relative radiometric error due to calibration drift.


                             We can write this relative radiometric accuracy, due to calibration
                          drift, as:

                                                ⎛ dG ⎞ 2  ⎛ dR ⎞ 2  ⎛ dG ⎞  2
                                      RA Drift  ≈  t ⎜  E ⎟ +⎜  D ⎟ +⎜  O ⎟   (10.51)
                                                ⎝  dt  ⎠  ⎝  dt  ⎠  ⎝  dt  ⎠

                          where      t =  time interval between the characterization and the
                                        object view measurement (s)
                                dG /dt =  relative rate of change of the electronic gain of the
                                  E
                                        system (Hz)
                                dR /dt =  relative rate of change of the responsivity of the
                                  D
                                        detector (Hz)
                                dG /dt =  relative rate of change of the system response due
                                  O
                                        to optical alignment (Hz)
                             This equation applies to a case where drift varies linearly with
                          respect to time (i.e., it always goes in the same direction). After a long
                          period of time, Eq. (10.32) will overestimate the radiometric error.
                          The three rates of change cited earlier are governed principally by the
                          change of temperature of the instrument.
                             Unlike other detectors, such as MCT units, InSb detectors exhibit
                          much less variation of responsivity with respect to temperature.
                          DR /dt is then negligible compared with other sources of calibra-
                             D
                          tion drift, whether the detector is cooled with a liquid-nitrogen
                          pour-filled Dewar or with a sterling cooler. This assumes, however,
                          that the temperature of the detector remains below a certain tem-
                          perature limit of about 120 K. In a mobile environment, where all
                          directions are permitted, special care must be taken to ensure that
                          none of the nitrogen is spilled if the Dewar is turned upside
                          down.
                             dG /dt is affected by temperature drift as well as by electromag-
                                E
                          netic disturbances, but these can be kept at a minimum through
   560   561   562   563   564   565   566   567   568   569   570