Improved edge detection and noise removal in gravity maps via the use of gravity gradients

Abstract

Standard edge detection algorithms perform poorly when applied to gravity maps of very high or very low resolution. Gravity maps of very high resolution, i.e., measured very close to the causative sources, are affected by noise and the resulting edges may be too cluttered for proper visual and numerical analysis. In gravity maps of very low resolution, i.e., measured very far from causative sources or calculated at high level of upward continuation, sources tend to interfere with one another and consequently edges of geological relevance may be impossible to detect. Algorithms specifically designed to detect edges and ridges in gravity gradient grids can help overcoming both problems. In low resolution gravity maps, they highlight features that are not detected by standard edge detection algorithms while on high resolution maps the combined use of gravity and gravity gradient grids can be employed to remove edges due to noise, vastly improving the visual inspection and geological interpretation.