High Performance Subpixel Hyperspectral Image Target Identification for Space-Based Sensors

Existing target identification technologies for space-based hyperspectral sensors do not attain the high detection and low false alarm rates required for effective battlefield awareness. The low and sub-pixel nature of targets in these images presents significant challenges to reliably recognize targets and separate them from surrounding clutter. In this effort, we propose to develop and deliver a phenomenology-guided adaptive statistical detector and supporting software for high performance subpixel HSI target identification. The proposed effort will build upon our Phase I success, where the SET / Johns Hopkins University Applied Physics Laboratory team has demonstrated novel approaches to primary stages of the processing chain including optimized background characterization, phenomenologically-based subpixel detection and adaptive threshold estimation that enable target/clutter separation that significantly surpasses existing technologies. The subpixel detector design incorporates physically meaningful constraints that enable enhanced detection while suppressing false alarms.  In Phase II, we propose to develop and deliver software modules for the Environment for Visualizing Images (ENVI) processing environment to enable individual module operation that may be combined with existing algorithms.  We will also deliver and evaluate the performance of an end-to-end HSI subpixel target detection system. BENEFIT: This technology will offer a new capability to enable high performance subpixel target identification for HSI data, enabling wide-area, space-based target identification not viable with current systems.  This technology has broad reaching applications for military and commercial applications including target identification, search and rescue operations and forensic investigations.