Micro-levelling of aeromagnetic data using a moving differential median filter

Abstract

Micro-levelling is a process in which small line errors (corrugations) are removed from a set of data. Most microlevelling routines involve filtering a grid of data to detect errors along lines. this often implies using a combi- nation of 2-D highpass and directional filtering in the wave number domain, and requires lines to be sub-parallel. An alternative micro-levelling technique is suggested which is based on correction of levelling errors using a moving differential median filter. The correction value for a point along a line of data is taken to be the different between median values for a 2-D and a 1-D filter window. The 1-D filter operates only on the line currently being processed while the 2-D filter window also includes data from neighbouring lines. The technique requires the removal of a regional field before it can be applied, but it allows lines to be non-parallel and irregular. The technique istested using aeromagnetic data recorded during a period of severe diurnal variations, and it is proven to be effective in removing residual line errors without affecting high wave number, geological anomalies.