Utilizing Optical Defocusing and Computational Imaging for Laser Intensity Reduction at CCD Focal Planes

Devising a means to reduce the fluence at the image sensor by a factor of 100 would eliminate the threat from all but the largest, cumbersome sources. We propose that the needed fluence reduction can be achieved by operating the existing imager at a strongly defocused setting, and applying hardware-accelerated existing state-of-the-art image deblurring algorithms to its data stream. Several existing deblurring algorithms are capable of very good inverse filtering performance with a 19x19 blur kernel which corresponds to an approximate fluence attenuation of 360 . The system concept is to interface the unmodified image sensor with a signal processor to implement the deblurring algorithm at the video rate. Aerodyne Research Inc (ARI), and our partner, the University of Notre Dame (UND) spin-off Emu Solutions offer innovative and patented techniques for FPGA hardware acceleration of "non-streamable" algorithms that typify inversion problems in computational imaging. Our Phase II goal is a full-up system implementation, coupling a representative optical imager to a real-time video-rate FPGA implementation of the computational deblurring. Our Phase I efforts will establish the feasibility of our approach, employing computational emulations, laboratory optical confirmation, and preliminary fixed-point arithmetic assessment.