High Performance Computing for DEAP Imaging through Atmospheric Turbulence

Previous research indicates the digital holography (DH) methods when combined with model-based iterative reconstruction can solve the challenging problems posed by the need to image through deep turbulence real time. In this Phase I STTR research, we propose to analyze and design a high-performance computing system and associated Deep-turbulence Estimator for Anisoplanatic Propogation (DEAP) algorithm for real-time imaging through turbulence. Key challenges will be to achieve the bandwidth and latency required for both ISR/imaging applications and wave front estimation for close-loop adaptation of optical systems while also meeting practical weight, power, size, and cost constraints. Our proposed Phase I approach is to study the computation and memory performance and requirements of CPU, GPU, FPGA systems, and use this information to design an optimized system that can meet Air Force requirements. Our goal is than to use the resulting design for implementation of an operational system as part of a potential Phase II effort.