DETECTION OF MILITARY SIGNIFICANT CHANGES IN EO/IR SENSOR DATA

The ultimate goal of this multi-phase effort is to develop a semiautomated IMINT processing system for EO and/or IR sensor data that is based on object-level change detection (OLCD) concepts. The objective of the Phase I program is to demonstrate the feasibility of object level change detection in the context of substantial reduction in false alarms. Our approach to object level change detection is a multi-stage and modular processing architecture. The general architecture is similar to the SAIP OLCD architecture (developed by AAEC) with the exception that we are adding (1) stronger emphasis on feature-based change detection, (2) a second layer of verification for missing and new objects, and (3) a model-based processing paradigm. To provide state-of-the-art EO/IR OLCD capability, AAEC has teamed with Hughes Aircraft Corporation (HAC) to include efficient/real-time model-based target recognition capabilites for the proposed effort.In the Phase I program our specific objective is to develop and evaluate three key processing modules: geolocation accuracy, consistancy of target cueing, and an initial spatial-based features for change evaluation. The results of the Phase I effort will provide the basis and solid foundation for the design, implementation and demonstration of the full process in chain.