David Dikeman (Lockheed Martin), Mr. Scot Seto (Lockheed Martin)
Keywords: NROC, Non-Resolved Object Characterization
Abstract:
Lockheed Martin Hawaii presents a novel signal processing algorithm for focal plane array processing. We introduce the Harmonic Structure Function (HSF) and demonstrate its capability in detecting, classifying and counting rotating bodies in a single pixel. The HSF has a powerful use in dynamical situations occurring on scales less than the single pixel solid angle. The work presented here is making a major impact in the Missile Defense Agencys Project Hercules Forward Based Sensor (FBS) group but the results presented here is shown in an unclassified form. First, the HSF algorithm is detailed. The origin of the HSF is in the ASW (AntiSubmarine Warfare) acoustic processing domain and the analogy to the focal plane is given. Next, the mathematical definition of the HSF and the natural extension from integral to discrete form is detailed. Thereafter, additional harmonic processing techniques such as the so-called sidelobe reduction are explained. These techniques are powerful methods to determine the fundamental frequency of a given rotating body that can have various harmonically related narrow band tonal structures. Simulations of rotating bodies and modulating reflectance used for analysis are then discussed. These simulations result in the construction of time series data for rotating bodies with fundamental frequencies in noisy backgrounds. The HSF is then used to analyze these fidelity simulations. It is shown that the HSF is capable of detecting, classifying and counting objects on a single pixel. Finally, the robustness of the algorithm is analyzed and it is shown that the number of detectable objects is dependent on sample rate, target temporal extent, and other factors. This analysis yield important considerations for sensor developers and operators.
Date of Conference: September 10-14, 2006
Track: Non-Resolved Object Characterization