A moving hum filter to suppress rotor noise in high-resolution airborne magnetic data

Abstract

A unique filtering approach is developed to eliminate helicopter rotor noise. It is designed to suppress harmonic noise from a rotor that varies slightly in amplitude, phase, and frequency and that contaminates aeromagnetic data. The filter provides a powerful harmonic noise-suppression tool for data acquired with modern large-dynamic-range recording systems. This three-step approach — polynomial fitting, bandpass filtering, and rotor-noise synthesis — significantly reduces rotor noise without altering the spectra of signals of interest. Two steps before hum filtering — polynomial fitting and bandpass filtering — are critical to accurately model the weak rotor noise. During rotor-noise synthesis, amplitude, phase, and frequency are determined. Data are processed segment by segment so that there is no limit on the length of data. The segment length changes dynamically along a line based on modeling results. Modeling the rotor noise is stable and efficient. Real-world data examples demonstrate that this method can suppress rotor noise by more than 95% when implemented in an aeromagnetic data-processing flow.