Detection and classification of underwater targets by magnetic gradiometry

SHORT ABSTRACT: This paper describes the theoretical analysis of a method to detect and classify objects from magnetic maps. The method, originally designed to detect buried UXO by land survey with a magnetometer, is adapted here for underwater survey with a gradiometer. Evaluation is performed on simulated data in order to predict the performance of the adapted algorithm to this new situation. We have simulated a gradiometer composed of three magnetometers that measure the intensity of the magnetic field. The paper presents results obtained by the classification algorithm using the gradient maps as input to locate the targets and estimate their depths and magnetic moments. Examples illustrating the influence of the sampling grid and errors in the location of the gradiometer underwater on the detection and classification performance are also presented.