Identification of Unexploded Ordnance from Clutter Using Neural Networks
Anna Szidarovszky*, Zonge Engineering and Research Organization, Tucson, Arizona;
Mary Poulton, University of Arizona, Department of Mining Engineering, Tucson, Arizona;
Scott C. MacInnes*, Zonge Engineering and Research Organization, Soldotna, Alaska
Paper – [pdf] UXO_IDfrClutter_NeuralNetworks
The largest costs associated with subsurface Unexploded Ordnance (UXO) remediation are associated with removing non-UXO. Discrimination between UXO and non-UXO is important for both cost and safety reasons. A neural network was developed to distinguish between UXO and non-UXO clutter using TEM data. There are two stages for the learning process of a neural network: training and validation. A synthetic dataset was created using actual acquisition configurations, with varying amounts of random noise. This dataset included 934 UXO targets representing seven different UXO types, and 789 clutter objects based on four templates with varying size and random asymmetry. The results show 97% accuracy for correctly classifying clutter, and
97% accuracy for correctly classifying UXO. The level of classification success is based on the classification Receiver Operating Characteristic (ROC) curves. The ROC curve represents the relationship between UXO classified correctly (Hit rate) versus clutter miss classified (False alarm).
In many countries around the world, UXO remain from warfare or from military practice. UXO pose a problem to society in two main ways: cost of detection and removal, and safety of civilians and removal crews. UXO are differentiated from landmines and this paper focuses solely on UXO.