Practical implementation of edge detection methods to reliably interpret the lunar gravity data with detailed comparative studies

Abstract Enhancing the ability to detect subsurface geological features via potential field (gravity and magnetic) data remains a key challenge. We conducted comprehensive comparisons to adapt the proposed Modified Horizontal Gradient Amplitude (MHGA) edge detector for robust interpretation of lunar gravity data from the GRAIL mission in the Gardner region. Initial comparisons were performed on synthetic gravity datasets, both noise-free and noise-contaminated, using MHGA and 11 other widely-used methods. MHGA outperformed others in terms of accuracy, resolution, sharpness, and amplitude balance of output for noise-free data. However, like the rest 11 methods, MHGA is significantly affected by noise amplification during the derivative calculation. We thus further examined various noise attenuation techniques, including the recent β -VDR and MNLM methods, in conjunction with these edge detectors. The impact of tuning parameters on the performance of these denoising techniques was thoroughly analyzed. Finally, the enhanced MHGA’s results for the Gardner region’s gravity data were validated against available 3-D inversion outcomes. The combined findings revealed a large, dense structure beneath the topographic bulge of the Gardner Plateau, likely representing the remnant magma reservoir that fed the region’s volcanic complex. These results underscore the effectiveness of MHGA, particularly when appropriately adapted, in addressing edge detection challenges. Additionally, the effects of applying different regional-residual separation methods to the lunar data were detailed and the discussion part shed light on the accuracy loss problem of the noise attenuation process, emphasizing the need for careful management of noise suppression to preserve important short-wavelength information.