Enhanced 3D Gravity Inversion Using ResU-Net with Density Logging Constraints: A Dual-Phase Training Approach

Gravity exploration has become an important geophysical method due to its low cost and high efficiency. With the rise of artificial intelligence, data-driven gravity inversion methods based on deep learning (DL) possess physical property recovery capabilities that conventional regularization methods lack. However, existing DL methods suffer from insufficient prior information constraints, which leads to inversion models with large data fitting errors and unreliable results. Moreover, the inversion results lack constraints and matching from other exploration methods, leading to results that may contradict known geological conditions.In this paper, we propose a novel approach to address the above issues. First, we introduce a depth-weighting function to the neural network (NN) and train it in the weighted density parameter domain. The NN, under the constraint of the weighted forward operator, demonstrates improved inversion performance, with the resulting inversion model exhibiting smaller data fitting errors. Next, we divide the entire network training into two phases: we first train a large pre-trained network Net-I, and then use the density logging information as the constraint to get the optimized fine-tuning network Net-II. Through testing and comparison, the inversion quality of our method has significantly improved compared to the unconstrained data-driven DL inversion method. Additionally, we also conduct a comparison and discussion of our method with the focusing inversion method. Finally, we apply this method to the measured data from the San Nicolas mining area in Mexico, and compare it with two recent gravity inversion methods based on DL.

====================================================================================

We have uploaded multiple large files to Quark Cloud. (such as "Supported_Data" folder, "Net_Model" folder, "Tra&val.zip" folder, and "Syn.rar" folder)

Please download them through this link: https://pan.quark.cn/s/7df7fb8635ce

The extraction code is evi6