The CoreSmart Predictor provides a cost-effective assessment of grades for most metals in
existing drill core archives
cores from ongoing drill campaigns (diamond drilling, RC or percussion drills)
material from surface sampling
hyperspectral data from drone and airborne missions
During development the AI has been tested on independent samples for the most important industrial metals. (see table).
The CoreSmart Predictor is a smart piece of Artificial Intelligence that has been trained on more than 1300 km hyperspectral core scan data and 130000 geochemical analyses.
This AI is a specially developed neural network that processes hyperspectral scan data (from VNIR, SWIR and/or TIR bands), achieving an accuracy for metal contents of between 85% and 95%. This enables you to extract more accurate information from your hyperspectral data.
You can use these results in near real-time and plan your geochemical analyses based on it.
While the system can be used for drill core analysis, airborne data and also in operational mining - the database behind the learning system is itself a vast pool of under-utilised data for strategic green field exploration. Metal grades of a wide variety of metals for thousands of drill holes are fully available based on the predictor in addition to assayed segments of the core. The use of this data and our expertise in complex data analysis and modelling opens-up new approaches for identifying potential industrial and rare metal targets. As the data pool is increasing and new metals are integrated into the CoreSmart predictor, the usage of proxies for REE and Lithium detection improves.
A validated and quality assured database containing hyperspectral scan data of more than 3000 drill holes and 700.000 assays from all Australian states and beyond
CoreSmart predictions for all segments of publicly available hyperspectral scanned drill holes in Australia
Tools for importing scanned drill core data from different sources