SPECTRAL SCIENCES, INC — Department of Defense SBIR Phase I: N202-118

Description

Unmanned aerial vehicles (UAVs) carrying surveillance equipment, targeting systems and even explosives and bombs, are a growing threat to US armed forces and government agencies.   This has in turn produced a need for systems to detect UAVs around naval ships, military bases, government and public facilities and national infrastructure.  Spectral imagers, operating in the Midwave Infrared (MWIR) and Longwave Infrared (LWIR) atmospheric windows, have significant potential to provide day/night detection and classification of UAVs under a variety of terrain and sky conditions and discrimination from potential interfering objects like birds.  For such complex environments, it is imperative that the overall design of the sensor, and especially the selection of spectral windows covered, and spatial, spectral, and temporal resolution, be paired with specialized algorithms for detection, tracking and classification in highly cluttered conditions.  Spectral Sciences, Inc. (SSI) has significant experience with both spectral imaging system design and development, and spectral detection algorithms.  For this SBIR effort, we propose to develop a set of algorithms that will be used to detect UAVs using their spatial, spectral and temporal characteristics, together with sensor hardware and software specifications for optimal UAV detection.   In Phase I, we will collect high resolution spectral imagery from a number of UAVs, both stationary and in flight, against clear sky and cluttered backgrounds, to clarify the challenges to spectral detection.  Beginning from state-of-the-art spectral anomaly detection and material classification algorithms, our detection and classification algorithms will address specific issues of UAV detection, including atmospheric effects, frame to frame motion, day/night detection, and clutter suppression.