IR sensors suffer from non responsive or bad pixels and detector nonuniformities which introduce a texture to the uncorrected output that can occupy a significant portion of the total dynamic range and superimpose a static or fixed grainy pattern to the uncorrected output. The so called fixed pattern noise (FPN), under low illumination conditions, can render the signal to noise of the scene contrast edges to FPN to be near to, and often less than, unity (as shown) and therefore must be corrected in order to produce useable output.

CSC has developed and tested prototype scene based NUC using our AIP ASICs onboard UAV platforms for the short wave infrared SWIR (InGaAs) and long wave infrared LWIR (microbolometer) spectral bands. One implementation using CSC's processors requires differencing the input signal with a quasi-static spatial high pass filtered representation of the input to remove slowly varying high spatial frequency content. This processing removes the static pixel-scale fixed pattern noise (FPN).