Understanding the phenomenology of polarimetric data is necessary if we want to obtain the maximum benefit when we exploit that data. To first order, polarimetric phenomenology is driven by two things; the target material type (specular or diffuse) and the illuminating source (point (sun) or extended (body emission)). Polarimetric phenomenology can then be broken into three basic categories; ([specular material/sun source], [diffuse/sun], [specular/body]) where we have assigned body emission to the IR passband where materials are generally specular. The task of interest determines the category of interest since the task determines the dominant target material and the illuminating source (eg detecting diffuse targets under trees in VNIR = [diffuse/sun] category). In this paper, a specific case study for the important [diffuse/sun] category will be presented. For the reflective regime (0.3 - 3.0μm), the largest polarimetric signal is obtained when the sun illuminates a significant portion of the material BRDF lobe. This naturally points us to problems whose primary target materials are diffuse since the BRDF lobe for specular materials is tiny (low probability of acquiring on the BRDF lobe) & glinty (high probability of saturating the sensor when on lobe). In this case study, we investigated signatures of solar illuminated diffuse paints acquired by a polarimetric hyperspectral sensor. We will discuss the acquisition, reduction and exploitation of that data, and use it to illustrate the primary characteristics of reflective polarimetric phenomenology.
展开▼
