Abstract:Based on strapdown marine gravity real data, we check the effectiveness of using wavelet threshold in filtering marine gravity anomaly. We test different wavelet, threshold techniques and decomposition levels. Results show that the db6, db7, db8, db9, db10, sym6, sym7, sym8, sym9, sym10, coif3, coif4 and coif5 are the thirteen wavelets suited for marine gravimetry, and wavelet decomposition level can be taken to 8 or 9 layers. The unbiased risk threshold filtering effect is better. When the wavelet is decomposed into the 8th and 9th layers, the result of the wavelet threshold filtering agrees well with the filtering results using the Butterworth low-pass filter with 0.005 Hz and 0.0033 Hz cut-off frequency, but the result of the wavelet threshold filtering is more smoothing than the Butterworth low-pass filter. Further, the RMSE of the difference between the two filtering results is within 0.25 mGal. Wavelet threshold filtering is easier to decompose the noise component than the Butterworth low-pass filter, which can eliminate gravity distortion more effectively.