Feature fusion-enhanced t-SNE image atlas for geophysical features discovery

Area of Science:

• Geophysics
• Geological Data Analysis
• Mineral Exploration

Background:

• Traditional geophysical data analysis tools struggle with complex geological structures and integrating diverse datasets.
• Existing methods often overlook subtle patterns and relationships across different geophysical data types.

Purpose of the Study:

• To develop a data-driven approach for integrating diverse gridded geophysical datasets.
• To autonomously reveal inherent geological patterns without predefined assumptions on feature geometry.
• To facilitate intuitive data navigation and exploration for earth scientists.

Main Methods:

• Utilizing Haralick texture descriptors to encode geophysical data patches into a unified high-dimensional space.
• Employing t-distributed stochastic neighbor embedding (t-SNE) for nonlinear projection into a 2D interactive space, creating a 't-SNE Atlas'.
• Integrating magnetic and gravity data from the Yilgarn Craton, Western Australia.

Main Results:

• The 't-SNE Atlas' allows intuitive exploration of complex relationships between geophysical responses and geological structures.
• The methodology successfully reveals subtle geophysical patterns and facilitates the discovery of new geological insights.
• Demonstrated application in the Yilgarn Craton highlights the potential for identifying mineral prospectivity.

Conclusions:

• The proposed two-step approach offers a powerful, adaptable tool for geoscientific exploration by integrating diverse gridded datasets.
• The 't-SNE Atlas' enhances the discovery of geological features and supports hypothesis formulation.
• This method serves as a valuable pre-planning tool for exploration geologists, potentially accelerating the identification of promising exploration sites.