Data Processing Method of Dynamic Gravity Gradiometer Based on Time-frequency Combination

Gravity gradiometer is an instrument for continuous measurement of small gravity gradient changes on the earth's surface. The gravity gradiometer based on the principle of rotating accelerometer is the only gravity gradiometer to realize near surface dynamic measurement. Due to the weakness of gravity gradient signal and the limitation of accelerometer performance, the direct output signal of gravity gradiometer contains a lot of motion noise under dynamic conditions, and the signal-to-noise ratio is very low. In conventional gravity gradient data processing, the noise outside the measurement frequency band is removed by low-pass filter after motion compensation, but the dynamic noise distribution is full frequency band, and the low-pass filter cannot eliminate the noise in the passband. In addition, the real gravity gradient signal also has the characteristics of multi-scale and non-stationary in space. Therefore, the time-domain filtering link of wavelet threshold filtering is added in data processing, and the limitation of Fourier transform to non-stationary signal is solved to a certain extent by using the multi-resolution analysis characteristics of wavelet transform in time domain. According to the characteristics of surface gravity gradient signal, the appropriate wavelet base is selected, and the influence of threshold control method, threshold estimation method and decomposition layers on the filtering effect is discussed in detail by using the measured data of gravity gradient vehicle test. On this basis, the optimal filtering parameters are determined. Finally, the above time-frequency combined data processing method is used to process the measured data of gravity gradient aerial test. The processing results show that compared with conventional low-pass filtering, the data processing method proposed in this paper can improve the internal coincidence accuracy of airborne gravity gradient measurement by 30% without losing spatial resolution, and has obvious dynamic noise suppression effect.