SBIR/STTR Award attributes
Physical Sciences, Inc. will develop an underwater mine imaging and detection capability based on a novel flash LADAR sensor employing compressive sensing (CS) and a single pixel architecture. The approach makes dual use of orbital angular momentum light as the projected measurement basis for CS and in the form of a vortex imaging receiver resulting in spatial separation between ballistic and scattered photons. These attributes will preserve spatial frequencies of mine-like targets and retrieval of signal photons from the target return that are up to two orders of magnitude below the background scatter level. The sensor is projected to achieve 10 cm range resolution out to 8 m for clear water (c = 0.1 m-1) and out to 2 m for highly turbid water (c = 2 m-1) with a 128×128 spatial format (256×256 with pan sharpening) over a 30º field of view. The capability is coupled with fast CS reconstruction engine using split Bregman iteration followed by an automated target recognition (ATR) algorithm based on a convolution neural network (CNN). A prototype CS flash LADAR and real time CS and CNN-based ATR algorithms will be developed and demonstrated.
