Deriving Nonlinearity Coefficients from Changing Background

Sensor nonlinearity is usually characterized with an external test target that is varied in intensity to change the radiance on the detector. A disadvantage of this technique is that the varying intensity of the test target must be accurately known. For a sensor such as the Cross-track Infrared Sounder (CrIS), which has a built in radiance calibration system, it is practical to use a constant temperature external test target and let the varying background change the radiance on the detector. CrIS is an infrared Fourier transform spectrometer used to make weather and climate observations. The CrIS detectors are linear but require a quadric nonlinearity correction to meet radiometric accuracy requirements. Each CrIS cross-track scan includes a measurement of deep space and an internal ICT (Internal Calibration Target) that floats with the sensor temperature. This calibration information is averaged, nonlinearity corrected, and used to continually update the calibration of the sensor. If the nonlinearity correction coefficients are correct, the calibrated radiance measurement of a constant temperature target will remain constant even with a changing sensor background. An iterative method was used to adjust the nonlinearity correction coefficients to minimize the change in calibrated radiance for a fixed temperature target under changing background conditions. These iteratively obtained nonlinearity coefficients agreed well with the coefficients determined with the conventional method of changing the target temperature. An advantage of this iterative technique is that neither the absolute radiance of the external target or the background is required. This paper will include a description of the algorithms used to derive the new linearity correction parameters as well as results found when applying linearity correction with both the traditional and new method.