AI-based targeting is becoming a useful addition to mineral exploration workflows, giving geoscientists a systematic way to evaluate complex project information and test geological ideas at scale. For this analysis to be effective, data needs to be prepared to support consistent spatial assessment across the area of interest (AOI).
In DORA, VRIFY’s AI prospectivity mapping software, Data Augmentation is a core data preparation step after standardization and unification are complete. Through a set of proprietary modules, this process supports three main functions: transforming vector inputs into rasters, enhancing signals in existing rasters, and, where appropriate, generating new layers that add geological and spatial context to model development.
Below is a summary of each function, with examples of specific modules to show how Data Augmentation prepares exploration data for modelling.
Dig further into the full DORA workflow and how its outputs remain interpretable in this article: Transparent AI for Mineral Exploration.
Transforming vectors to rasters
For projects that include vector data, Data Augmentation converts these inputs into rasters DORA can ingest. Structural linework and field measurements for faults, veins, and shear zones are common examples, as are surface geochemical points from rock, soil, till, or lake sediment sampling.
The gridded output is generated at a resolution guided by input density. Sparse source data produces a coarser grid, while denser coverage can support finer spatial detail.
Once converted, these layers become part of the data stack used for subsequent modelling. Rasterization also provides geoscientists with a continuous spatial representation of the source observations, making relationships between features easier to interpret across the AOI.
Enhancing existing rasters
Some project datasets, like magnetic grids, are already in raster form and can be used for modelling in DORA without conversion. A source raster may carry useful information, but the feature of interest is not always explicit in the measured response. A magnetic response, for example, may reflect alteration corridors, lithological contacts, or faults even when these features are not mapped directly. Data Augmentation processes existing rasters to extract patterns in the measurements. These enhanced rasters can be added as additional input layers alongside the source rasters, giving DORA more information to draw on during modelling.
Generating new data layers
Beyond format transformation and enhancement, Data Augmentation can generate new layers from one or more raster inputs. These derivatives extend what can be evaluated from the source data by applying known spatial or geological patterns across the AOI, including areas where direct measurements are limited.
Derived layers should be introduced to modelling with a specific test in mind. A practical workflow starts with objective datasets that have continuous coverage, after which derived layers can be added progressively when a geological feature or signal needs clearer representation. Comparing model response with and without a derived layer helps show whether its inclusion strengthens the result or adds uncertainty from interpretation-derived information.
<hr />
Data Augmentation prepares source inputs for modelling in DORA by rasterizing vector data and enhancing signals present in source rasters. New layers generated through this process can broaden the analytical context, but should be selected according to the deposit concept and the geoscientist’s understanding of local controls on mineralization rather than carried forward by default. As new data is acquired and the project’s geological picture develops, Data Augmentation can be iterated upon to refine the model.
If you would like to learn more about Data Augmentation or DORA, book a demo with our Geoscience Team.
.png)
.png){kind=spacer}
.jpg)
.png)
