The main objective of this project is to determine feasibility of using one-pass, dynamic underwater (UW) electromagnetic induction (EMI) data sets for UW targets detection, mapping, and identifications. Specifically, the project’s objectives are as follows:

  • Process Ultra transient electromagnetic Array (UltraTEMA) system’s data sets collected at the Sequim Bay test site in Washington State using the advanced EMI models and signal processing approaches.
  • Study how the sensor motion effect on UW targets EMI responses and use these understandings to improve targets detection and classification.
  • Extract responding magnetic dipoles polarization tensors and locations from each dynamic data point; cluster the extracted locations in the global coordinate system; determine detected targets (informed source) location, and group primary, secondary and tertiary extracted magnetic polarizabilities inside each cluster.
  • Assess and document the applicability of the one pass EMI systems and advanced data processing approaches for UW targets detection and classification using different EMI data sets.

Technical Approach

This research investigates applicability and limitations of the advanced EMI systems and models for detection and identification unexploded ordnances (UXO) targets in aquatic environment by processing the UltraTEMA system’s EMI data sets collected at the Sequim Bay test site in Washington State. The UltraTEMA system consists of four transmitter coils and twelve vector receiver sensors. Thus, the system measures complete EMI response of a target at each dynamic data point and provides 144 data value at each nth (n=1,2, .., 25) time gate. To take advantage of these reach data sets, each dynamic data will be processed, and targets intrinsic and extrinsic parameters will be extracted and classified using the advanced EMI models. Namely, first background response will be removed from the raw UW electromagnetic data using high pass filter. An optimal length of the high pass filter will be research and determine. Next, each dynamic data point will be inverted via the advanced EMI models, such as the orthonormalized volume magnetic source (ONVMS) and differential evolution algorithms, and targets intrinsic, such as effective magnetic polarizability (i.e. total ONVMS), and extrinsic locations will be extracted. The algorithms will account for sensor motions as well as for signals due to marine environment. Third, the extracted extrinsic parameters will cluster and be used for locating UW anomalies. Finally, the estimated intrinsic parameters will be used to classify anomalies as target of interest and clutter via the library matching and statistical classification approaches.


This project will determine the feasibility of using advanced EMI sensors in dynamic mode for UW UXO targets detection and classification. If successful, the project will pave the way toward operational dynamic one pass EMI systems for UW unexploded munitions detection and classification.