Wavelet analysis for the regional-residual and local separation of potential field anomalies

A method based on the discrete wavelet transform was applied to the regional-residual separation of potential fields and to the filtering of local anomalies. A specific space-scale wavelet analysis, called multiresolution analysis, allowed decomposition of the signal with respect to a vast range of scales. Different analysing wavelets were applied to anomalies in both synthetic and real cases, but the more appropriate basis needed to be chosen by requiring the maximum compactness for the multiresolution analysis. Moreover, since such analysis was found not to be shift-invariant, the same criterion was applied to choosing the best signal shift. Application of the technique to both synthetic and real cases produced an optimal space-scale representation of the fields and a consistent regional-residual separation. Furthermore, the space localization allowed a variety of filtered signals to be obtained, each one with a specific scale and local area content. Fourier and wavelet analyses were both applied to the filtering out of the intense Etna anomaly from the aeromagnetic field of Sicily. The wavelet method was more powerful, suppressing only the Etna volcano anomaly and leaving the rest of the map practically unchanged.