Sensors that can capture several frames per second of hyperspectral image data have the potential to significantly improve the capability of persistent surveillance systems. In this project we will exploit this opportunity by developing a system INV-PS that combines motion constraints with spectral and spatial information for target detection and tracking. The system will learn adaptive target models from time sequences of spectral and high-resolution image data. The use of general adaptive models will support use with various sensors, viewing geometries, and target and background types. The new approach is designed to allow the tracking of a large set of targets moving in close proximity in complex urban environments under changing illumination conditions. The Phase II project will generate a high-speed implementation of INV-PS that can perform real-time tracking on the sensor platform. The new models and algorithms will be assessed over a large set of data. A detailed commercialization plan is given for the new software.
Benefits: Due to its unique position in the marketplace, INV-PS has great potential for the DoD and Intelligence Community. The new product will greatly improve the operational utility of systems for a range of intelligence, surveillance, and reconnaissance applications. INV-PS also has a high likelihood of success for Commercial/Civil applications in areas such as traffic monitoring and border monitoring. HyperTech is already involved in several programs that will lead to commercialization opportunities for INV-PS.
Keywords: Hyperspectral, Remote Sensing, Target Detection, Target Tracking, Persistent Surveillance, Motion, Real-Time, Identification
