2022 Q3 technical highlights

 

This edition features the following technical highlights:
•Characterising Regolith Interfaces in Drill Holes in NSW; and
• Automated 3D Modelling

Please feel free to share this quarterly update amongst your networks.

 

CHARACTERISING REGOLITH INTERFACES IN DRILL HOLES IN NSW

Program 3 Researchers have developed a semi-automated, objective workflow for characterizing cover materials and identifying key interfaces within the cover using data that can be collected at the drill site or core yard in near real time.

The majority of exploration drill holes will intersect regolith, which can either hide or reveal indicators of mineralisation at depth. The ability to recognise facies interfaces or other discontinuities within the cover is an essential first step to deciphering geochemical signals in the cover rocks and understanding their potential as pathfinders to mineralization – either in the cover or in the underlying basement. Unfortunately, the recognition of cover interfaces in drilling samples can be difficult, even for experienced geologists, and highly subjective.

Our workflow involves collection of drill core geochemical and mineralogical data (for example portable XRF, handheld SWIR and VNIR spectrometers or HyLoggerTM) which are processed using multivariate classification and optimized Machine Learning (ML) algorithms. Selected data products (typically ratios, scalars and indices based on multiple data inputs) are used to generate multiscale classification and identification of interfaces using the Data MosaicTM software. The Data MosaicTM output closely resembles a geological log, but is data driven and objective. This delivers reliability and consistency and often uncovers subtle but important signals that are not identified by conventional geological logging procedures. Preliminary results indicate that measurements taken using field portable instruments at the drill site or core yard, at low cost and in near real-time (e.g. pXRF, handheld spectrometers) deliver comparable results to laboratory measurements (e.g. lab assays, HyLoggerTM).

The current study was based on three diamond drillholes from the Cobar NDI campaign areas in central western NSW. However, the proposed workflow is independent of geography or geology. In the future we will apply, and refine, the workflow based on legacy data, legacy samples and new drilling samples collected during National Drilling Initiative campaigns.

 

AUTOMATED 3D MODELLING

Program 2 researchers are currently developing QGIS plugins which will provide broad and free access to the automated 3D modelling and inversion algorithms developed by MinEx CRC as part of the Loop3D consortium.

One of the aims of MinEx CRC is to facilitate rapid 3D model building to aid geological understanding, guide exploration targeting and ultimately facilitate discovery in areas which have not yet been explored in detail. Loop3D has delivered a battery of new 3D modelling and inversion algorithms. However, although these algorithms are readily via the Loop3D website, they require specialist technical knowledge to be put to effective use.

The QGIS plugins will facilitate access by non-specialist users on a familiar (and free) software platform. This will make it easier to generate, visualize and interrogate 3D geological models in the context of other exploration datasets within the same GIS platform.

The QGIS plugins are based on algorithms previously developed by MinEx CRC as part of the Loop3D consortium. For example, the first pass of geological polygon and polyline layers will be extracted from the GIS package and processed for input into a 3D model using our “loop-processor” plugin, which uses built-in dictionaries developed for the stand-alone “map2loop” software.

The QGIS plugins will be made available for download as they become available via the Loop3D github repository. Users will be able to download the plugins, modify and use as they wish. Our hope is that this will enable MinEx CRC participants access to the software in a user-friendly format and encourage the exploration community to participate in further development of the software.