Extended Representation of Mueller Matrices

Abstract

The so-called arbitrary decomposition of a given Mueller matrix into a convex sum of nondepolarizing constituents provides a general framework for parallel decompositions of polarimetric interactions. Even though arbitrary decomposition can be performed through an infinite number of sets of components, the nature of such components is subject to certain restrictions which limit the interpretation of the Mueller matrix in terms of simple configurations. In this communication, a new approach based on the addition of some portion of a perfect depolarizer before the parallel decomposition is introduced, leading to a set of three components which depend, respectively, on the first column, the first row, and the remaining 3 × 3 submatrix of the original Mueller matrix, so that those components inherit, in a decoupled manner, the polarizance vector, the diattenuation vector, and the combined complementary polarimetric information on depolarization and retardance.