Underground Face Mapping

Mapping faces is a critical process; understanding the geology up close and guiding production. It's the highest level of detail you're likely to encounter and not being restricted to the diameter of an HQ core can bring out the creative side when it comes to capturing it. Hole logging is heavily standardised to ensure any of the dozens of geos passing through a project are (broadly) recording the same things, the same geology codes, have the same understanding of alteration intensity and log to the same level of detail. I remember as an exploration geo thinking a project would be easy; logging basalt, sandstone, siltstone and coal over a section of about 50m, until being told we were to log 1 cm scale changes of lithology. Enforcing the same uniformity to face mapping is harder still, but particularly when pencils and paper come into play.

Tablet based face mapping is one way of addressing this and some other issues with a paper based mapping process. It offers:

• Improved efficiency
• Realtime Data Capture
• Accuracy
• Standardisation
• Data Integration and Sharing
• Better Collaboration

Challenges to consider before implementing,

• Learning Curve
• Upfront Costs for tablets and software
• Battery Dependency

I've made an example for comparison on the dataset above focussing on just the drawing part, which is only actually one part of the larger process. An extract of a typical hand drawn face map is shown, a lot of detailed lithology has been recorded, some faults mapped and a channel sample position has been marked in the face. This would have been conducted rapidly on an A4 sheet of paper. What isn't displayed is a table filled in by hand, showing the from-to of the samples and later the grades of those samples. The approx. area of the geology represented in the samples is calculated for an area based face grade estimate. Significant structures are noted and their orientation recorded in another table and a small image showing the position of this face and structures is sketched out. The name of the closest survey station and the distance is recorded on the sheet, along with the name of the geo, the level, the drive name, date and a channel sample ID. Ultimately the calculated grade is used to determine the destination and this is written on the sheet and with any of the geological observations.

All of this information then has to be shared, so the face map is scanned and the samples are typed into some form of spreadsheet and ultimately uploaded to a central database. The scanned face map is registered to its true position and stored likely on the server. These hundreds of slightly different face maps are called upon in areas of model updates, where the senior resource geo traces contacts onto the image to make their model align with the reality.

Simple steps and a well trodden process, but a lot of work and duplicated effort. The results are also not available immediately.

Improvement can be made by replicating the process above, but in a digitised way. The results of that are shown in the second and third image. In this case, all of the information captured in the paper mapping sheet is captured by using a tablet. The drawing of the geology and the channel sample position is in it's true 3D position first time, meaning no need to register and trace over the mapping. The samples are directly saved into an SQL database along with the mapping itself. This makes it instantly available and recallable, giving immediate spatial context to the observations. Photography can also be taken with the table (assuming good lighting) and registered easily if needed to assist in mapping or allow mapping to be conducted later.

Coming back to Power BI, having the sampling and mapping stored in an SQL database, it's a short step away from using that data and sharing it more widely through the use of dashboards. If there is wifi at the face, it could be basically live.

The next evolution would be scanning of a face - getting away from the mapping onto a diagram of unfolded boxes, towards mapping very accurately and directly onto a scan of the face. Structural observations could be pulled out with more detail and a lot more of them could be made than might be possible without a scan. The quality of image and scan obtained from a scanner with a good lighting rig means observations could (gulp) be made without needing to visit the face at all. The scans would be useful to multiple stakeholders beyond just the grade control geologist, before and after scans could highlight shotcrete thickness issues, bolting problems and accurate mined volumes among other things.