Polarimetric scene modeling in the thermal infrared

Interest in polarimetric remote sensing is gaining momentum in the visible and remains strong in the microwave regions of the spectrum. However, passive polarimetric phenomenology in the 3-14 micron infrared (IR) region is complicated by the relative contributions and complementary polarization orientation of the thermally emitted and background reflected radiance. Although this modality has found success in specific missions (i.e. surface-laid landmine and tripwire detection), the dependence on time of day, scene conditions, scene geometry, collection geometry, etc. makes it difficult to easily perform empirical instrument design or tasking trade studies. This paper presents improvements to the modeling framework within the Digital Imaging and Remote Sensing Image Generation (DIRSIG) model to polarimetrically render scenes in the infrared. The DIRSIG model rigorously treats the polarimetric nature of both thermally emitted and background reflected scene radiance. The correct modeling of these two components is key to accurately predicting polarized signatures for various instrument designs and collection scenarios. The DIRSIG polarized BRDF and polarized directional emissivity models are described and compared to experimentally measured data. Results showing the sensitivity of polarimetric IR phenomenology to target and background material properties, collection geometry, and scene configuration are presented.