Gravity gradient signal extraction based on time-frequency feature threshold method

Abstract

Gravity gradient, defined as the spatial derivative of gravitational acceleration, represents the detailed variation of the gravity field and plays a significant role in resource exploration and auxiliary navigation. Airborne gravity gradient measurement is an important method for obtaining high-precision gravity gradient data. Compared to traditional gravity measurement techniques, it offers superior measurement efficiency and provides high-frequency information of gravity field which reveals small-scale terrain features. However, it also introduces more high-frequency measurement noise. Due to the randomness of Earth's topography and resource distribution, the high-frequency characteristics of gravity gradient signals are mainly concentrated in specific time and frequency ranges, exhibiting distinct time-frequency features. This paper employs the VMD and Hilbert transform to obtains the time-frequency features of gravity gradient signal, and proposes a time-frequency feature threshold method to extract gravity gradient signal. This method can accurately analyze and extract the gravity gradient signals components without being affected by the same frequency noise, suppressing 78 % of the noise in simulation data. Additionally, the proposed method is validated in the gravity gradient data of actual terrain, demonstrating over 19 % improvement compared to traditional smooth filter and wavelet threshold methods. Furthermore, it effectively retains the high-frequency characteristics of gravity gradient signal, enhancing the advantages of airborne gravity gradient signals