Implicit neural representations for 3D gravity inversion

doi:10.1016/j.cageo.2025.106082

Implicit neural representations for 3D gravity inversion

2025-12-02

0

PRE

AI

DOI：10.1016/j.cageo.2025.106082

Original

Original request for help

Save

Abstract

En 中文

• An implicit gravity inversion method is developed, the governing physical laws of gravity is embedded into the network optimization. • The proposed method circumvents the need for large training datasets and eliminates the reliance on manually designed regularization terms. • Explored the potential source of implicit regularization effects from the network architecture. • Evaluated on synthetic and the San Nicolas field dataset, the method achieves fidelity reconstructions and exhibits robust performance.

Implicit neural representations for 3D gravity inversion

Abstract

Journal

Organization