Automated Feature/Anomaly Extraction from Synthetic Aperture Radar (SAR) Coherent Change Detection (CCD) Imagery

SAR imaging is a key capability for Army programs such as Future Combat Systems and Aerial Common Sensor. Coherent change detection (CCD) is a powerful technique to examine large SAR images and identify areas where enemy activities have occurred. However, continual changes occur in populated regions making it necessary to differentiate military and civilian activities. To address this challenge, TSC will a develop a suite of innovative real-time algorithms that can segment the terrain into homogeneous regions, identify areas in which changes have occurred, extract features such as straight line and curved vehicle tracks, classify vehicles based on the track separation, detect foot traffic, locate areas that have been used for military activity (e.g. resupply, bivouac) and locate areas being prepared for military use (e.g. revetments, launch sites). A key feature of TSC's approach is to use not only a CCD image, but also the multiple reference images that are provided by persistent surveillance. The CCD imagery will be differenced to remove areas of low correlation, eliminate persistent false alarms, and detect soil subsidence. Another key feature of TSC's approach is establishing normalcy models to help differentiate routine and anomalous activities.