Detection and identification of underwater (UW) unexploded ordnance (UXO) is a significant problem in remediation efforts to reclaim formerly used defense sites. Presently there is no system available that can accurately survey and map the location of UXO at UW sites and reliably discriminate UXO from UW debris. The UW UXO survey problem is complicated by a wide range of targets (e.g., magnetic and non-magnetic, metallic and non-metallic) that are often buried. Sensor detection concepts are currently being developed for UW UXO detection and mapping including multi-sensor systems capable of independently measuring different target physical characteristics. UW detection sensors and systems tested to date have had limited measurement capability versus buried UW targets and there is no set of quality data from independent detection sensors to assess the value of new detection and identification sensors and algorithms.
The objective of the Multi-Sensor Data Base (MSDB) project was to collect a data base of high quality sensor data from multiple sensors on a range of UW buried UXO targets deployed in shallow water (10-15 feet) at a variety of environmental sites and to perform data fusion analysis of independent sensor target detections to assess the improvement in UXO target location and discrimination of multiple sensor data versus single sensor data.
The MSDB was collected using a set of state-of-the-art acoustic, passive magnetic and active electromagnetic (EM) sensors deployed from a catamaran in two UW towed vehicles at positions above or on the sea bottom within detection range of each sensor. The UW sensor platforms were deployed in fixed positions with respect to the catamaran to permit co-registration of the sensor responses during the search operations.
Phase I MSDB tests conducted in October 2005 included over 75 search runs over the UXO target fields at two shallow water sites in St. Andrews Bay, Florida. Phase I tests demonstrated that the acoustic (Buried Object Scanning Sonar [BOSS]) and EM sensors (Realtime Gradiometer and GEM-3 array) could detect proud and buried UXO-sized targets at both sites, but only the BOSS had sufficient detection range to reliably detect UXO targets during a single search track run. Phase I tests also showed that a wide area augmentation system (WAAS) global positioning system (GPS) was not suitable for detailed target mapping and did not permit meaningful fusion of the UXO detection data from the three different sensors. Phase II tests were conducted in September 2008 with a modified goal of collecting quality detection and mapping UW UXO with only a BOSS sensor using a Real-Time Kinematic (RTK) GPS with higher mapping accuracy. Phase II tests demonstrated that the BOSS acoustic sensor could detect and map all UW UXO targets within 25 cm of ground truth and reliably detect UXO targets buried 30 cm below the sand bottom.
MSDB data is available to permit study and evaluation of UW UXO sensor detection techniques and UW UXO sensor phenomenology that can lead to successful approaches to UW UXO mapping.