Automatic contact detection in side-scan sonar data

Abstract

Side-scan sonar is a proven tool for detection of underwater objects, particularly those objects that project above the seafloor. Rapid assessment of side-scan imagery for object detection is critical for port security needs. However, current side-scan data processing techniques are largely manual, highly time-consuming, and prone to operator error. Availability of well-trained analysts is also a challenge. This article describes a research and development effort at Science Applications International Corporation to automate side-scan sonar contact detection for safety of navigation surveys. Included in the development effort are innovative image processing and machine learning techniques designed to reduce the number of false alarms. These automated techniques are directly applicable to port security operations.