EVMS is a scientific Python framework to reconstruct a 3D subsurface source-intensity field from surface gamma measurements. It combines a sparse forward model, geological regularization, and an interactive Streamlit interface for reproducible inversion workflows.
- Sparse, scalable inversion on voxel grids (
scipy.sparse, iterative solvers). - Physics-informed forward model with attenuation and distance kernel.
- Geological priors through layer-aware smoothing and finite fracture barriers.
- Publication-ready outputs: volumetric field export and textured mesh export.
- Interactive analysis, diagnostics, and trust reporting in Streamlit.
Forward model:
Voxel discretization:
Regularized inversion:
where
conda env create -f environment.yml
conda activate evms-env
pip install -e .streamlit run streamlit_app.pyThe app provides:
- Inversion workspace
- Method page (equations and workflow)
- Credits page
- Measurements CSV:
x,y,z,value - Grid:
.npymask or.objmesh (voxelized in-app) - Fractures JSON: optional finite rectangular fracture patches
- Calibration CSV: optional
x,y,z,cps/s
pytestThe test suite covers geometry, forward operator behavior, regularization, inversion, calibration, and diagnostics.
Documentation and project updates: maxsc4.github.io
Maxime SOARES CORREIA — soarescorreia@ipgp.fr
Copy-paste citation:
SOARES CORREIA, M. (2026). EVMS: Volumetric Inversion of Geological Radioactivity from Surface Gamma Measurements (v0.1.1) [Computer software]. GEOPS — Géosciences Paris-Saclay (UMR 8148, CNRS), Université Paris-Saclay. https://doi.org/10.5281/zenodo.18586941