"Despite its sensitivity to diverse shape, surface, and material properties, laser polarimetry is one of the primary remaining remote-sensing modalities yet to be effectively utilized. The traditional limitations of laser polarimetry in a field sensor have been 1) making efficient use of multi-dimensional polarimeter data, and 2) achieving required sensor footprint and speed. Phase 1 demonstrated automated target recognition (ATR) algorithms based on mission-specific laser-polarimeter data that achieve significant performance improvements over algorithms based on non-polarized laser data. In particular, these algorithms achieve target pose invariance and low false-alarm rate (FAR). These algorithms, generally termed polarization-components techniques (PCT), overcome the first limitation to fielding a polarimetric ATR sensor. The PCT algorithms also minimize measurement requirements, thereby guiding the design of partial Mueller matrix polarimeters (pMMP), which represent a significant step in overcoming the second limitation. The Phase 2 program will finalize and build a conceptual pMMP design and apply the resulting field polarimeter to mission-specific field tests. Future commercialization efforts will be supported by a field polarimeter available for demonstrations of general-purpose ATR and by the next generation PCT test and simulation software to be developed and utilized in the Phase 2 program."
Keywords: "laser Polarimetry, Machine-Learning, Mueller-Matrix, Support Vector Machine, Atr, Far"
