Research on Adaptive Filtering Methods for Shipborne Gravimetry

The results of shipborne gravity measurements contain a variety of noises because of the complexity and variability of the marine environment. Therefore, one of the key technologies to increase the accuracy of measurement results is high-performance filters. In response to this issue, this article incorporates the Singular Spectrum Analysis (SSA) method into shipborne gravity noise reduction processing. Meanwhile, an improved Variational Modal Decomposition (D-VMD) method is proposed, which determines the optimal number of decomposition layers based on the moving speed of the gravimeter (m/s) and sampling frequency, and the optimal IMF component is determined based on the sample entropy of each IMF component, thus obtaining the optimal signal for noise reduction processing for shipborne gravity. Finally, using the data from a dgM1 gravimeter, the noise reduction capabilities of the SSA, D-VMD, empirical mode decomposition (EMD), and variational mode decomposition (VMD) methods were compared and examined. The results demonstrated that the proposed D-VMD method achieves comparable denoising accuracy with SSA. On the tested profiles, D-VMD yields small-magnitude accuracy improvements over conventional VMD and EMD. Its main advantage lies in the adaptive initialization strategy, which reduces the manual parameter-tuning burden for shipborne gravity data processing.