Unlocking Geotechnical Insight with AI-Driven 3D Vision Technology
From automated mine work face mapping to size distribution analysis, our solutions aim to transform how the industry interprets and leverages geotechnical data. With our technology, decision-makers gain deeper insight faster, enabling safer operations, reduced costs, and better resource estimation.
Automated Mine Work Face Mapping
We applied our fracture detection algorithm to real-world open-pit mine images, and our system accurately identified the mine work face, capturing precise angle and size information, this information was then projected onto a fully rendered 3D model of the open-pit mine.
Rock Fracture Characterisation Using Borehole Data
Our custom-built hardware and software system was tested in the lab, it combines 3D modeling with AI to accurately capture rock fractures. The system not only identified fracture details with precision but also projected them directly onto 3D models, delivering clear, visual insights into core sample — reliable, easy to set up, and ready for real-world application.
Size Distribution Analysis from Images
Our rock fragment recognition algorithm was put to the test in both lab and real mining environments - and it delivered. In controlled lab settings, it accurately measured rock sizes with results nearly identical to physical sieving tests, outperforming other algorithms. On actual mine sites, it handled medium and large fragments with ease, and with just 30 minutes of quick tuning, it adapted to detect smaller rocks with precision. Fast, accurate, and fully automated, our algorithm is a powerful solution for rock fragment analysis in any environment.
Mineral Segmentation in Rock CT Images
Our team developed a deep learning algorithm for segmenting minerals in high-resolution rock CT scans. By incorporating structural cues into the model’s training process, the algorithm achieved higher accuracy and significantly improved preservation of mineral shapes and fine details. This advancement supports more precise and automated mineral identification, enhancing digital core analysis and geotechnical evaluations.
AI-Enhanced Lithology Classification Across Boreholes
SeeRock developed a simple yet effective data adaptation method to improve lithology classification from geophysical logs across unlabelled boreholes. By aligning log quantiles between reference and target holes, the approach reduces domain variability and boosts model transferability. Tested on 11 field boreholes using five log types, the method cut accuracy loss by up to 27% and improved lithology proportion match by up to 24%. This lightweight solution enhances regional lithological mapping and supports more consistent geotechnical modelling.
Our Valuable Customers
Who We Are
SeeRock was founded by a team of academic experts who recognised the challenges faced during traditional mining operations. With SeeRock, we aim to improve mining efficiency and safety at a more affordable price point.
Learn more