Abstract
Gravity data edge detection methods play a crucial role in identifying the horizontal positions of buried sources and enhancing the interpretation of subsurface structures. Among these methods, the Total Horizontal Gradient (THG) filter is widely used due to its noise stability and straightforward formulation. However, the THG filter has inherent limitations, prompting the development of refined edge detection techniques. To address these shortcomings, various approaches based on local phase analysis or normalized filters have been introduced, many of which incorporate the arctangent function to combine horizontal and vertical gradients. While these methods improve edge detection, they also exhibit drawbacks, such as the generation of spurious edges and restricted resolution. In this study, we propose the Modified Arctangent Function (MAT), which enhances gravity source edge detection by integrating the total horizontal gradient with a modified arctangent function. The effectiveness of the MAT filter is systematically evaluated against conventional filters that utilize the arctangent function and/or the total horizontal gradient. Its performance is validated through synthetic gravity data tests, both with and without noise contamination. To further assess its applicability, the MAT filter is applied to high-resolution gravity data from the WGM2012 (World Gravity Map) over the Rafsanjan Plain in Iran. Additionally, an approximated derivative calculation method is incorporated to mitigate noise amplification during the derivative computation process. The results from both synthetic and real data confirm that the MAT filter effectively detects and delineates gravity anomalies, demonstrating its potential as a valuable tool for geophysical interpretation.