Automated Change Detection: Applications for Synthetic Aperture Sonar and Future Capabilities

We present an overview of an automated end-to-end change-detection process as it is applied to synthetic aperture sonar (SAS) imagery. Although we are discussing an example of a single application, the automated change-detection (ACD) process can be applied to different imaging modalities, such as medical imaging, synthetic aperture radar (SAR), remote sensing, and radio astronomy. This article introduces the different processing steps of the automation: image coregistration, change-map generation, detection, and data reduction (false-alarm reduction) as it applies to automatic target recognition using sonar imagery. We examine ACD applications to multitemporal SAS imagery in detecting changes on the sea floor. In addition, we discuss some of the challenges and necessary approaches to augment the ACD process for operator decision-aid tools and performance assessment of the process.