🧭 GeoRad3D

3D Modelling and Visualization of Natural Radioactivity in Geological Outcrops

Figure — Example of 3D kriging and colorized geological mesh (GeoRad3D).

---

🧪 Project Overview

GeoRad3D is a scientific research project dedicated to the three-dimensional modelling of natural radioactivity within geological outcrops. It aims to bridge the gap between surface radiometric measurements and the internal distribution of radionuclides in the subsurface, using geostatistical interpolation and 3D visualization techniques.

The approach combines field data, geostatistics, and 3D graphics to generate volumetric models of radioactivity, offering a new way to explore, validate, and visualize radiometric signatures within geological structures.

Under the supervision of B. Saint-Bezar and A. Benedicto (GEOPS, Université Paris-Saclay).

---

🎯 Scientific Objectives

• Model the 3D distribution of radioactivity in rock outcrops using surface data.
• Extrapolate in depth (beneath the outcrop surface) using 3D ordinary kriging, accounting for spatial correlations and variograms fitted to the data.
• Quantify uncertainty through kriging variance and provide validation metrics (ME, RMSE, MSSE, VSE). (WIP)
• Visualize and explore volumetric radioactivity in 3D (isosurfaces, slices, point data overlay).
• Integrate geological meshes (.obj models) to map radioactivity directly on 3D geometry via vertex coloring or texture baking.

The resulting models can be explored interactively, exported, and compared with geological interpretations, providing a physically consistent and reproducible framework for studying spatial heterogeneities of natural radioelements within rocks.

---

⚙️ Methodology Overview

1. Data Preparation

   • Import and cleaning of georeferenced radiometric points (X, Y, Z, R), where R is the measured radioactivity (in counts per minute).
   • Automatic computation of a bounding box surrounding the dataset.
   • Creation of a regular 3D grid inside this domain.

2. 3D Interpolation (Kriging)

   • Ordinary Kriging in three dimensions using PyKrige (OrdinaryKriging3D).
   • User-selectable variogram model (spherical, exponential, gaussian, or linear).
   • Computation of both estimated values and kriging variance over the full 3D grid.

3. Cross-Validation (optional)

   • Leave-One-Out (LOO) validation to evaluate predictive performance.
   • Statistical indicators automatically computed:
     • Mean Error (ME)
     • Root Mean Square Error (RMSE)
     • Mean Standardized Squared Error (MSSE)
     • Variance of Standardized Errors (VSE)

4. Visualization and Exploration

   • 3D interactive visualization of the interpolated volume:
     • Isosurfaces of radioactivity levels (semi-transparent solids)
     • Neon-colored measurement points
     • Mouse-over tooltips and click-labels with values
   • 2D depth slices linked to a slider for exploring different Z-levels.
   • Gaussian smoothing and percentile-based color scaling for perceptual readability.

   (see CLI – view for command usage)

5. Integration with 3D Meshes

   • Ability to load external .obj geological models (with .mtl if available).
   • “Painting” of the mesh surface according to interpolated radioactivity values:
     • Vertex coloring (RGB per vertex)
     • or Texture baking (UV → PNG + updated MTL).
   • The resulting mesh can be directly visualized in any 3D viewer or GIS.

   (see CLI – paint)

---

📘 Documentation

For a detailed guide on how to use the GeoRad3D command-line interface (CLI),
see the CLI Documentation.

This includes explanations and usage examples for:

• 3D interpolation (fit)
• Volume visualization (view)
• Mesh painting (paint)
• Z-slice animation (gif)
• (soon) Model alignment (align-obj)

---

💡 Expected Outcomes

GeoRad3D provides a workflow to:

• Build spatially coherent 3D models of natural radioactivity;
• Evaluate the internal structure of anomalies within rock masses;
• Facilitate visual and quantitative comparison between radiometric data and geological features;
• Produce publication-ready figures and 3D exports for further interpretation or integration in GIS or modeling platforms.

By grounding surface radioactivity measurements into a volumetric and geostatistical context, the project contributes to a better understanding of radionuclide distribution processes in fractured and altered rocks, and to more realistic radiological mapping at the outcrop scale.

---

🏛️ Institutional Context

GeoRad3D is part of a research program conducted at GEOPS (Géosciences Paris-Saclay, Université Paris-Saclay).

• Supervisors:

  • B. Saint-Bezar (GEOPS)
  • A. Benedicto (GEOPS)

• Keywords: geostatistics, kriging, radioactivity, 3D modeling, geological outcrop, visualization