Abstract

Recent investigations in the detection and measurement of the linear polarization characteristics of a scene have suggested that polarization-sensitive sensors offer increased capability (over conventional sensors) in: (1) distinguishing targets (polarized) from a cluttered background (unpolarized), (2) discriminating between targets with different polarization signatures, and/or (3) establishing the relative orientation of a target. However, the detection/measurement of the linear polarization descriptors of a scene (degree of linear polarization, Stokes Vector, etc.) involves a nontrivial, arithmetic combination of a specific set of intensity measurements—each measurement with inherent accuracy/precision errors. The precision of the derived polarization parameters is seen to be a complex function of the precision of the individual intensity measurements and the specific measurement technique employed. This paper examines the relationship between measurement precision error [in terms of signal-to-noise ratio (SNR)] and the precision of two key derived polarization parameters (degree of linear polarization and linear polarization angle). The relationship is quantified for three widely accepted polarization measurement techniques (Pickering, modified Pickering, and Fessenkov methods). This analysis provides the system designer with a solid reference for determining the required SNR to achieve the polarization measurement precision necessary to meet system requirements.

© 1991 Optical Society of America