The design and construction of wide FOV imaging polarimeters for use in atmospheric remote sensing requires significant attention to the prevention of artificial polarization induced by the optical elements. Surface, coatings, and angles of incidence throughout the system must be carefully designed in order to minimize these artifacts because the remaining instrumental bias polarization is the main factor which drives the final polarimetric accuracy of the system. In this work, we present a detailed evaluation and analysis to explore the possibility of retrieving the initial polarization state of the light traveling through a generic system that has inherent instrumental polarization. Our case is a wide FOV lens and a splitter device. In particular, we chose as splitter device a Philips-type prism, because it is able to divide the signal in 3 independent channels that could be simultaneously analyze to retrieve the three first elements of the Stoke vector (in atmospheric applications the elliptical polarization can be neglected [1]). The Philips-type configuration is a versatile, compact and robust prism device that is typically used in three color camera systems. It has been used in some commercial polarimetric cameras which do not claim high accuracy polarization measurements [2]. With this work, we address the accuracy of our polarization inversion and measurements made with the Philips-type beam divider.
