Edge Effect Correction in the S-A Method for Geochemical Anomaly Separation

Anomaly separation using geochemical data often involves operations in the frequency domain, such as filtering and reducing noise/signal ratios.Unfortunately, the abrupt edge truncation of an image along edges and holes (with missing data) often causes frequency distribution distortion in the frequency domain.For example, bright strips are commonly seen in frequency distribution when using a Fourier transform.Such edge effect distortion may affect information extraction results; sometimes severely, depending on the edge abruptness of the image.Traditionally, edge effects are reduced by smoothing the image boundary prior to applying a Fourier transform.Zero-padding is one of the most commonly used smoothing methods.This simple method can reduce the edge effect to some degree but still distorts the image in some cases.Moreover, due to the complexity of geoscience images, which can include irregular shapes and holes with missing data, zero-padding does not always give satisfactory results.This paper proposes the use of decay functions to handle edge effects when extracting information from geoscience images.As an application, this method has been used in a newly developed multifractal method (S-A) for separating geochemical anomalies from background patterns.A geochemical dataset chosen from a mineral district in Nova Scotia, Canada was used to validate the method. KEY WORDS: edge effect correction, fractal modeling, spatial information extraction, zero-padding, decay functions.