Gravity inversion provides a fundamental tool for determining the subsurface density distribution, which is further used to delineate mineral bodies and explore the structure of the solid earth. High-resolution density models often entail numerous model parameters, resulting in significant memory demands and prolonged computation times. To address these challenges while preserving resolution, we propose a novel method based on the DEXP (depth to extreme points) imaging technique to reduce meshing space. This technique provides an approximate density distribution image, which not only helps determine the 3D meshing space but also allows for generating a mesh flag to constrain the meshing process for inversion. Evaluations with synthetic datasets reveal, at most, a 92% reduction in model parameters, from 32,000 for the whole mesh volume to 2,651 for the optimized mesh volume. Inversion results indicate a nearly 95% decrease in computation time. The method’s practical applicability is further demonstrated through real-world gravity data from Dida Jilin Province, Northeast China, where it successfully reconstructs mineral bodies between depths 80 m and 200 m with a memory reduction of 50%. These findings underscore the effectiveness of optimized meshing in reducing model parameters, conserving memory, and minimizing time consumption during the inversion process.
陈 et al. (Thu,) studied this question.