Nonuniformity correction using the constant-statistics constraint: Analog and digital implementations

Current infra-red focal point arrays (IRFPAs) are limited by their inability to calibrate out component variations. Typically, off-board digital calibration is used to correct nonuniformities in these detector arrays; special calibration images are used to calibrate the system at startup. One-time calibration procedures such as these do not take into account other operating points and will fail to recalibrate for any drift in the parameters. Using clues from neurobiological adaptation, we have developed the constant-statistics (CS) algorithm for nonuniformity correction of IRFPAs. Gain and offset variations are successfully calibrated using simple assumptions of the scene under view. We give results for calibration of 1D and 2D images using a digital implementation. We also show that the constant-statistics algorithm compares favorably to an existing LMS-based nonuniformity correction algorithm by Scribner(1) in terms of convergence rate and computational complexity. Finally, we review the results of analog circuitry that was designed and fabricated with a 2um CMOS technology. Measured results from our test-chip show that the system achieves invariance to gain and offset variations of the input signal. This hardware is targeted for eventual use for in-and behind-the focal plane implementations.