Improving Accuracy in Locating Magnetic Sources: Combing Homogeneous Locator and Extreme Values for Inversion

Abstract:

In this study, we present an algorithm to improve the accuracy in locating magnetic sources, especially dipole magnetic sources using a combination of a homogeneous locator and positions of extreme values. Homogeneous locator is a homogeneous function and point based algorithm that uses magnetic field intensity and its first and second order tensors. If using all magnetic data points for inversion of the locations of magnetic dipole sources, more solutions than actual targets will be produced. Also, interference from neighboring magnetic sources and noise will deteriorate the situation and make the locating task even more difficult. To improve such a situation, we investigated and compared various cases with different levels of interference and noise, and using a combination of the homogeneous locator and the positions of extreme values for inversion. Results show that (1) if the interference and noise levels are low, using magnetic and its first and second tensor values at the positions where the first vertical derivative of the magnetic field has extreme values – either maximum or minimum as inputs to the homogeneous locator resulted in the best results – accurate horizontal and vertical locations and structure indices; (2) if the interference and noise level are high, using magnetic and its first and second tensor values at the positions where the magnetic field intensity has extreme values as inputs to the homogeneous locator resulted in the best horizontal locations, but still using the values at the positions where the first vertical derivative of the magnetic field has extreme values produce the best vertical locations and structure indices. We applied the verified scheme to the field data for UXO detection and aeromagnetic data of Minnesota for geological structure study and the results are compared to Euler deconvolution and sole homogeneous locator method.