SBIR-STTR Award

Current systems that observe and detect changes from information gathered by a single sensor have limited success in dynamic scenarios. The fusion of information from desperate sources such as SAR and IR in conjunction with Object Level Change Detection offers the potential for improved performance in terms of reduced false alarms. This is primarily due to the complementary information across the two sensing domains and the differences in the source of false alarms. The incorporation of model-based processing into the processing chain will provide the capability to accommodate normal signature variations in sensor, and object masking effects. The overall goal of the proposed effort is to develop a robust process that detects changes at the object level. The developed change detection process will fuse IR and SAR sensor data at the feature level and will also provide the mechanism to incorporate other imagery sources including spectral data. It will complement ongoing fusion technologies currently developed under the DDB program. The proposed fusion will be performed at multiple stages of the exploitation chain and will take advantage of mode-based processing to improve system robustness. The feasibility of the developed technology will be demonstrated using sample data. We will show the potential for improved false alarm reduction over current capabilities.