The decision to use a fixed camera gain as opposed to using
either the camera hardware auto gain function or software to pro-
vide an automated gain function was based on a number of factors.
First, it proved difficult to consistently identify the putative origin
of a detected anomaly in an image when the brightness of that type
of anomaly varied in images due to automated changes in camera
gain. Second, automated gain adjustments frequently hindered vi-
sual detection of anomalies, particularly when the scene being im-
aged was dark or contained a very bright area. Third, the finite time
delay in adjusting gains sometimes resulted in a positive feedback
loop where the user would move the camera head in responses to
the effects of the changing gain, which would then initiate a new
round of gain change. Fourth, the time delay associated with wait-
ing for automated gains to stabilize increased effective scanning
times and times for acquiring hyperspectral data cubes. A single,
selectable, gain setting proved to be impractical because of varia-