Imagers of reflected light operate in the spectral band between 0.4 and 2.0 μm. The visible band is from 0.4 to 0.75 μm. The near-infrared (NIR) band extends from 0.75 to about 1.0 μm. However, it is not uncommon to call the whole region between 0.75 and 2.5 μm the near infrared. Light with wavelengths between 1 and 2 μm is called short-wave infrared (SWIR).
Evaluation of reflected-light imagers differs from evaluation of thermal imagers in two important respects. First, the target signature is highly variable. Both the intensity and spectral nature of natural illumination depend on time of day and weather. Also, there is a wide variation in material spectral reflectivities. An object that is highly visible when viewed against one background surface might almost disappear when viewed against a different surface material. The combination of varying illuminations and different background characteristics changes target signal by orders of magnitude.
Photo-electron counting models are generally used in the reflective bands. This is the second difference between reflective and thermal models. In the thermal model, both signal and noise are represented by a joules-per-second flux. Noise is represented by the watts on the detector that generates the same rms amplitude over a 1-sec time period. In the reflective model, the photon flux is integrated for a time period equal to a frame time or an eye integration time. Both models are accurate when properly applied. However, the different treatments lead to different calibration constants.
Section 10.1 discusses factors that affect target-to-background signatures. Section 10.2 develops the system contrast threshold function (CTFsys) for reflected-light imagers. Calculating imager resolution using the TTP metric is described in Section 10.3.