A correlation-based approach for determining the threshold value of singular value decomposition filtering for potential field data denoising

Abstract

We present a correlation coefficient analysis (CCA) method for obtaining threshold when using singular value decomposition (SVD) filtering method to reduce noise in potential field data. Before computation of correlation coefficients, SVD is performed on the gridded potential field data with the purpose of obtaining singular values of the data. A sliding window is utilized to truncate the acquired singular values, which allows us to obtain different singular value sequences. The lower limit of the sliding window is generally set to zero and the upper limit of the sliding window is the threshold. Then, we calculate and plot the correlation coefficients associated with the initial sequence and the newly obtained sequences, choosing the inflection point of the plotted correlation coefficients as the threshold. The CCA method offers a quantitative way to determine a threshold, which can be easily implemented by a computer program. We illustrate the method using synthetic datasets and field data from a metallic deposit area in the middle-lower reaches of the Yangtze River in China. The results show that the proposed method is effective and is able to provide an optimal threshold.