A global leader in diamond exploration and mining faced significant limitations in the manual analysis of drill core samples. This process relied heavily on expert petrologists to interpret complex hyperspectral imagery, introducing bottlenecks in throughput and consistency. As the organization pushed toward digital transformation in mining intelligence, the need for scalable, repeatable, and automated analysis became critical. The core challenge was to extract and classify spatial mineralogical features—such as mineral types, texture, layering, and particle distribution—from high-dimensional hyperspectral data.
Manual workflows limited the speed of exploration cycles and depended on scarce human expertise. The mining company engaged Opti-Num to automate these processes using artificial intelligence and image processing tools. Adding to the complexity were the technical uncertainties around preprocessing and modelling such large, complex datasets. The effectiveness of AI techniques could only be determined after thorough exploration of the data and close collaboration with domain experts.
Opti-Num adopted a structured, phased approach to reduce risk and deliver value incrementally:
Phase 1 – Exploration and Feasibility
We aligned closely with the client’s geoscientists to define objectives, assess hyperspectral core imagery, and identify high-value use cases. A proof-of-concept was developed to demonstrate the viability of using AI to automate core analysis.
Phase 2 – Implementation and Scale-Up
Building on Phase 1, we expanded the methodology to classify a wider range of features across a broader dataset. Machine learning models were trained using labelled data from expert petrologists and validated against real-world samples. Custom MATLAB applications were developed to support internal workflows and ensure usability by non-technical geoscientists..
Phase 3 – Enablement and Future-Proofing
To ensure long-term sustainability, we conducted targeted training sessions for the client’s staff, provided detailed documentation, and supported knowledge transfer. Models were integrated into internal systems, allowing the organization to independently operate, refine, and scale the solution.
Key technical highlights:
The successful automation of drill core analysis delivered significant strategic value to the mining organization:
Through close collaboration and a deeply iterative approach, Opti-Num helped a major mining organization modernize a traditionally manual process—showcasing how artificial intelligence can unlock deeper insights and drive smarter exploration in the mining sector.
.png)
.png)
