EdgeDetectPFI: An algorithm for automatic edge detection in potential field anomaly images – application to dike-like magnetic structures

Abstract

We propose an algorithm to automatically locate the spatial position of anomalies in potential field images, which can be used to estimate the depth and width of causative sources. The magnetic anomaly is firstly enhanced using an edge detection filter based on a simple transformation (the Signum transform) which retains only the signs of the anomalous field. The theoretical edge positions can be recognized from the locations where one of the spatial field derivatives (or a function of them) change its sign: the zero crossover points. These points are easily identified from the Signum transformed spatial derivatives. The actual sources depths and widths are then estimated using the widths of the positive plateaus obtained from two different Signum transformed data: one based on the vertical derivative and the other using the vertical derivative minus the absolute value of the horizontal derivative. Our algorithm finds these widths in an automatic fashion by computing the radius of the largest circles inside the positive plateaus. Numerical experiments with synthetic data show that the proposed approach provides reliable estimates for the target parameters. Additional testing is carried out with aeromagnetic data from Santa Catarina, Southern Brazil, and the resulting parameter maps are compared with Euler deconvolution.