An impediment to the success of unexploded ordnance (UXO) magnetic field surveys is the interference of target magnetic anomalies by background fields, including overlapping fields of neighboring shallow-sourced anomalies. The standard procedure is to apply a pre-inversion filter that attenuates some components of the background field, but leaves the target anomalies without appreciable distortion so that they can be input to full point-dipole inversion. Pre-inversion filtering, especially the de-median filter, is a valuable tool, but unfortunately, without distortion of the target anomalies, significant background fields and overlapping anomalies inevitably remain to hamper inversion or even detection of target anomalies due to UXO or clutter.

The objective of this project was to develop high-pass intra-inversion filtering (IIF) as a new method to improve inversion of magnetic anomalies in UXO magnetic field surveys. IIF allows accurate inversion of UXO magnetic anomalies in the presence of low-frequency background fields and overlapping anomalies of neighboring UXO and clutter.

Technical Approach

IIF, as developed for this project, takes the approach that a target anomaly can be severely distorted yet the filtered magnetic field data can be input to full point-dipole inversion with accurate recovery of all of the point-dipole parameters that are valuable to the analysis that follows. It accomplishes this task by high-pass filtering with a digital filter and then applying that exact same filter to the test models’ fields used internal to the inversion algorithm. In the case of the least-squares inversion scheme used in this project, the IIF is applied to the fields of three coincident unit-dipoles at each possible depth for all-node inversions or at each possible dipole location for flag-node inversions.


IIF was applied to anomalies in the Blind Test Area of the UXO Standardized Test Site located at Aberdeen Proving Ground, Maryland, wherein clutter and UXO (actually inert ordnance) were emplaced at various depths and inclinations. The results demonstrate that IIF significantly improves the accuracy of inversions of magnetic anomalies in UXO magnetic field surveys.

Researchers also unexpectedly invented the edge-adaptive gapped gradient-nulling (EAGGN) filter. This new filter automatically accommodates data gaps, survey edges and corners, flag-node data, single swaths of towed-array data, and restrictions of the filter input to any desired data window.


UXO cleanup costs are reduced by proper characterization of magnetic sources to distinguish UXO from some non-UXO sources. Accurate and efficient inversion of magnetic survey data ultimately will save excavation costs.