Accurate calibration of polarimetric sensors is critical to reducing and analyzing phenomenology data, producing
uniform polarimetric imagery for deployable sensors, and ensuring predictable performance of polarimetric algorithms.
It is desirable to develop a standard calibration method, including verification reporting, in order to increase credibility
with customers and foster communication and understanding within the polarimetric community. This paper seeks to
facilitate discussions within the community on arriving at such standards.
Both the calibration and verification methods presented here are performed easily with common polarimetric equipment,
and are applicable to visible and infrared systems with either partial Stokes or full Stokes sensitivity. The calibration
procedure has been used on infrared and visible polarimetric imagers over a six year period, and resulting imagery has
been presented previously at conferences and workshops.
The proposed calibration method involves the familiar calculation of the polarimetric data reduction matrix by
measuring the polarimeter's response to a set of input Stokes vectors. With this method, however, linear combinations
of Stokes vectors are used to generate highly accurate input states. This allows the direct measurement of all system
effects, in contrast with fitting modeled calibration parameters to measured data. This direct measurement of the data
reduction matrix allows higher order effects that are difficult to model to be discovered and corrected for in calibration.
This paper begins with a detailed tutorial on the proposed calibration and verification reporting methods. Example
results are then presented for a LWIR rotating half-wave retarder polarimeter.