Utilization of Full Tensor Eigenvalue and Eigenvector Analysis on Magnetic Gradiometer Data to detect UXO Targets

The detection of unexploded ordnances (UXO) targets based on a magnetic gradiometer system, has been widely researched in past decades. In this study, a novel UXO detection algorithm is suggested, which is based on full tensor eigenvalue and eigenvector analysis on magnetic gradiometer data. Firstly, the physical procedure of tensor eigenvalue and eigenvector decomposition for a UXO magnetic data is analyzed in detail. Then, a UXO detection algorithm is proposed based on the properties of the magnetic tensor eigenvalue and eigenvector decomposition. This algorithm can automatically locate magnetic the centroids of dipole-like sources without estimating the magnetic moment direction, or the effective magnetization direction. Finally, the effectiveness of this detection algorithm has been tested with the designed synthetic data. And an air-borne magnetic gradient data is utilized to validate the performance of the proposed algorithm. All of the experiments show that the algorithm is suitable for locating weak, anomalous dipole-like UXO targets and air-borne magnetic gradiometer data interpretation.

An estimated million acres of land in worldwide are contaminated with various types of UXO. To clean up these threats, air-borne magnetic surveys is a most common geophysical methods used to detect these metallic targets. There are many automatic detection methods for using total magnetic intensity (TMI) data. Recent studies show that magnetic gradiometry can significantly prove the signal to noise ratio (SNR) of magnetic measurement, and magnetic full tensor data has been increasingly used in wide area assessments to rapidly locate potential UXO contamination. And many corresponding methods have been proposed. Although these algorithms can efficiently detect the profile of the UXO, the centroids of the UXO source cannot be directly located in previous research. So, in this study, we suggest an algorithm that estimate the centroid of UXO source directly. This study is based on the physical properties of eigenvalue and eigenvector decomposition on magnetic full tensor data. Although eigenvalue analysis on curvature gravity gradient tensor (CGGT) has been widely discussed, the algorithm that is integrated the eigenvalue and eigenvector decomposition on full tensor magnetic data has not been presented.