Spectral Analysis of Magnetometer Swing in High-Resolution UAV-borne Aeromagnetic Surveys

Abstract

Electromagnetic interference produced by multirotor UAVs has the potential to compromise the integrity of UAV-borne total magnetic field (TMI) measurements collected with high-resolution optically pumped magnetometers. One method to overcome this challenge involves suspending the magnetometer sensor below the zone of electromagnetic interference via a semi-rigid mount. The semi-rigid mount allows the magnetometer payload to freely move in the pitch and roll axes, while rigidly fixing the yaw of the magnetometer to that of the multi-rotor UAV. The swinging motions of the magnetometer suspended below the UAV have the potential to introduce periodic variations in the collected magnetic field data. Within this study, spectral analysis was applied to UAVborne TMI measurements to assess contributions to the signal from the swinging, semi-rigidly mounted magnetometer payload. Overall, it was concluded that when the magnetometer was placed outside the zone of electromagnetic interference created by the UAV, compensation and filtering was not required to achieve industry standard measurements. This result was due to the magnetometer swinging through the relatively low-amplitude geomagnetic field gradient. However, when the magnetometer was placed within the zone of UAVinduced electromagnetic interference, a periodic, highfrequency signal was apparent in the TMI measurements. This was determined to be caused by the swinging of the suspended magnetometer payload within the high-gradient electromagnetic field produced by the multi-rotor UAV. The periodic signal (~0.35 Hz) was successfully identified and removed with a low-pass filter in the frequency domain, resulting in TMI measurements of industry standard quality. Filtering is a necessary step to avoid contaminating the magnetic field signals originating from sub-surface targets with unwanted signals related to the swinging of the magnetometer. Filtering can be applied when the targeted signal frequencies and the swinging signal frequencies do not spectrally overlap. This relationship must be considered in order to avoid removing important target signals during the filtering process.