Low Cost Multi-Spectral CS Sensor for Instantaneous Target Identification and Tracking

SBIR Phase I proposal proposes to develop a low cost autonomous multi-spectral compressive sensing (CS) seeker/sensor with machine learning (ML) and machine vision (MV) capability. The sensor enables wide area situational awareness, narrow field of view for detail observation and real-time target identification and tracking capabilities. The sensor consists of a wide field of view (WFOV) camera in spectrum l1, a narrow FOV (NFOV) observation camera in spectrum l2 for target details, another NFOV camera for target identification and tracking in spectra l3 and l4. The WFOV camera and the NFOV observation camera are conventional fast frame rate (e.g., 1 kHz) high resolution (e.g., 2k´2k) 2-D cameras.  The NFOV identification & tracking camera is the hyperspectral CS sensor in, e.g., mid-wave and long-wave IR bands where 2-D cameras are prohibitively expensive. An assistant vision camera is synched with the WFOV camera to align the NFOV cameras to the tracked target via the Saccade mirror pair that optically shift the pupil so as to pan & tilt the view like a gimbal. In Phase I, a preliminary design, trade analysis and model validation on a sensor design will be conducted for prototype development in Phase II which reaches TRL of 6.