Tag Archives: drill core

Tasmania drill core library receives investment boost

The Tasmania Government says it is expanding its geotechnical testing capabilities, with a A$2.4 million ($1.9 million) upgrade to its Mineral Resources Tasmania Core Library at Mornington.

The upgrade, expected to be completed by the middle of the year, will help the Australian island state both retain and grow its natural resources work, Guy Barnett, Minister for Resources in the Tasmania Government, expects.

Local construction company Fairbrother has been awarded the contract to upgrade the library, which will combine laboratory facilities, currently spread across two sites, and provide an upgraded and expanded state-of-the-art facility for geoscientific and analytical functions, which will better serve both government and industry needs, Barnett said. It will also provide an up-to-date interface for Mineral Resources Tasmania’s engagement with industry.

“This is a significant investment in a sector that supports more than 5,100 direct job, contributes more than 51% of our state’s exports, and produces product with a value of more than A$2 billion each year,” he said. “When our resources sector is strong, our economy thrives, and that is why we are making a significant investment into the scientific capability available right here in Tasmania.”

This is the first major upgrade since the library was first opened some 30 years ago, according to Barnett. The Core Library already stores more than 770 km of drill core and around 70,000 rock samples from across the state.

“This facility is a vital resource for our mining, exploration, research and education, and broader industrial sectors, and the upgrade will make it more efficient, effective and accessible to industry,” Barnett added. “This upgrade will play a significant role in realising our mineral potential and supports our collaborative efforts in working with industry through our existing scientific and exploration support packages.”

Technology symposium set to uncover new mineral exploration techniques

An eclectic mix of topics including the use of bacterial DNA to determine mineral deposits, hyperspectral imaging of core samples and the rise of quantitative data analysis will be discussed at a two-day conference presented by IMDEX.

The IMDEX Xploration Tech Symposium, usually held in Vancouver, Canada, will this year be held online on January 12-13 featuring a range of international speakers. The conference brings together experts at the forefront of innovation in the mining and exploration industries and will examine the latest in new technologies, tools, and advanced analytics, the mining tech company says.

Among the speakers will be Chad Hewson, Manager, Geophysics and Innovation, at Teck Resources; Dr Ralf Tappert, Co-Founder of Hyperspectral Intelligence; and Dr Thomas Bissig, a consultant geoscientist with over 25 years of experience.

Dr Bissig and colleague Bianca Phillips, a PhD student, will discuss the use of unconventional techniques for geochemical exploration including bacterial DNA, gases and selectively sampling areas where seismic pumping may have occurred to determine potential sub-surface deposits.

“(Bacterial DNA) is a technique that really only has become a possibility with increased computing power and lower costs of genomic sequencing,” Dr Bissig said. “It’s a great example of interdisciplinary research. The biologists and the geologists can work together to find deposits.”

Dr Bissig added: “We’re still learning how bacteria specifically respond to the geochemistry. An example would be bacteria that eat sulphides for their energy source; if you have sulphides in the ground which typically are associated with elements of interest that we would like to mine, we can detect sulphides in DNA.”

He said early studies in Canada’s Northwest Territories had returned “compelling signals, much better than conventional grid sampling of soils”.

Seismic pumping is the assessment of elements brought to the surface in water during seismic events. This is where Dr Tappert’s focus on developing a robust, portable hyperspectral scanning tool for examining core samples comes in, with the result being the geoLOGr rock analyser (pictured). Hyperspectral Intelligence has sold geoLOGR units to mining companies in Canada and South Africa.

“Spectroscopy is relatively simple technology,” Dr Tappert said. “You just have to put it into the right instrument and make it usable for mines and exploration companies.”

He said companies were putting more effort into collecting basic data.

“The drill core logging is the basic information that the entire mine relies upon, especially with deposits where you’ll have marginal grades. It really depends on the accuracy of your basic geological model to make the mine feasible or not.

“Companies have realised this is important, and spectroscopy plays a key role. It’s essentially the only method that you can use to collect continuous compositional information from the drill core.”

Hewson’s presentation, meanwhile, will examine the move from qualitative to quantitative data interpretation using existing and emerging technologies.

“That could be from field-portable tools or drilling technology, whether it be in-hole or in the core shack, and then methods which will transform the data into quantitative geoscience products,” he said.

Datarock machine learning drill core analysis tool hits major milestone

DiUS and Solve Geosolutions have leveraged PyTorch-based image analysis techniques to help automate the analysis of drill core imagery and provide greater insight that can be used to influence decision making on mine sites.

Mine sites often produce between 100 and 1,000 m of drill core per day, generating hundreds of images a week at a single drill site, according to PyTorch, an open source machine learning platform.

Historically, these images have been kept as a record of the job and a resource for geologists to refer to, rather than being used as a quantitative dataset that adds value to a mining operation.

Tapping into this rarely used data source of drill core imagery, DiUS – an Australia-based technology services organisation with a strong focus on machine learning and deep learning image segmentation analysis – joined forces with Solve Geosolutions – a mining-focused data science and machine learning consultancy – to build a machine learning-powered, cloud-based platform to automate the analysis of this drill core imagery using image segmentation technology.

Together, DiUS and Solve Geosolutions worked on applying a range of PyTorch-based image analysis techniques, including image classification, object detection and both semantic and instance segmentation, to a range of geological problems.

In particular, the team wanted to understand how different models performed in terms of training and inference speed, training requirements and overall accuracy of prediction to inform how they could be deployed in a production environment.

One model they used extensively is Mask R-CNN.

“This model can be applied to a range of segmentation tasks, however it can also demand large training datasets that are sometimes not available,” PyTorch said. “To support this, the team developed novel ways to increase the initial, often sparse training dataset through data augmentation techniques such as rotation, flipping, contrast, saturation, lighting and cropping.”

Following the initial discovery period, the team set about combining techniques to create an image processing workflow for drill core imagery. This involved developing a series of deep learning models that could process raw images into a structured format and segment the important geological information, according to PyTorch.

Their first productionised process was a metric referred to as RQD, otherwise known as rock quality designation. “RQD is a difficult and monotonous dataset to collect manually,” PyTorch said. “It’s also well suited to automation, and of high value to a mining operation. RQD is used by engineers to understand the strength of a rock and is used in the design and engineering of a mine.”

With the release of Detectron2 – a PyTorch-based computer vision library released by Facebook in October 2019 – the team made the decision to switch from the previous model implementation on TensorFlow to the next-generation platform to help improve instance segmentation tasks, PyTorch said.

The team found Detectron2 to be four times faster in training the models (using GPUs) and three times faster in inference (using CPUs) than the previous model implementation, PyTorch said.

Building the models on PyTorch-based frameworks meant the team was able to reduce valuable training time across the board. This increased the number of experiments and, as a consequence, improved model accuracy on an identical dataset. The PyTorch Dynamic Graph also made it much easier for the team to debug and investigate any issues that arose, PyTorch said.

The resultant Datarock platform is a software as a service offering that applies machine learning – image segmentation technologies – to drill core imagery and delivers information about a mineral deposit’s geology at scale, and at a resolution that’s not been previously economically viable, PyTorch says.

Since launching Datarock in 2019, the team has extended the platform to turn drill core imagery into high-quality datasets to support decision making throughout the entire mining cycle.

“The models perform optical character recognition, instance and semantic segmentation, as well as geological statistical analysis on a dataset,” PyTorch said. “This allows a geologist to inspect the model prediction and check for quantity and quality in unmatched datasets.”

Mining and exploration companies can now get consistent geological information from their rock core imagery in a matter of minutes, according to PyTorch.

“This near real-time power is enabling more intelligent decisions to be made further down the mining chain – saving time and money that can be put towards other business-critical projects – and freeing up geologists to do higher value tasks,” PyTorch said.

To date, the Datarock platform has processed more than 1 million metres of drill core images – that’s enough core to cover the distance between Sydney and Melbourne – over 800 km.

MICROMINE’s Whitehouse to explore machine learning in exploration at APCOM 2019

MICROMINE says machine learning has the potential to transform mineral exploration, and the company’s Ian Whitehouse intends to discuss just how at the upcoming APCOM 2019 conference, in Poland.

More than 500 delegates from across the globe are expected to travel to Wroclaw, in June, to discover the latest developments in the application of technology in the mineral industry at the 39th Application of Computers and Operations Research in the Mineral Industry (APCOM) conference.

Whitehouse, MICROMINE’s Geobank Product Strategy Manager, will be a keynote speaker at the symposium, which has the theme “Mining Goes Digital”.

Whitehouse said the application of machine learning to the process of collecting and analysing geological data in mineral exploration has the “potential to transform the way explorers operate”. He will delve into just how during his “Transforming Exploration Data Through Machine Learning” presentation on June 6.

“By adding machine learning to the process of collecting and analysing geological data, we vastly reduce the time a geologist spends doing administration work, enabling more time to concentrate on the quality and analysis of the data collected,” he said.

“This type of offering creates opportunity to lower exploration costs and increase the amount of data that can be collected, which are key drivers of the mining industry and will contribute to more exploration projects being approved.”

The traditional process of plan – drill – observe – measure – analyse, can be inefficient, and the application of technology and machine learning can address common issues such as inconsistent data collection and categorisation, Whitehouse said.

“In the exploration industry it is very common to find that one geologist has classified a rock and the next has classified it as something different. This has huge complications when trying to model the data. However, machine-learning algorithms can be used to fix these inconsistencies and errors in the databases prior to the resource geologist working with the data.”

Machine learning can be tapped by the resources industry to streamline geological processes, such as cleansing and validating data prior to starting the modelling process, according to MICROMINE.

Whitehouse said high quality DSLR cameras can provide a tool for exploration companies to collect high-quality imagery of core and chip trays, with machine-learning algorithms able to recognise features in the images.

“It is feasible for this data to be automatically collected and stored in a database,” he said.

To illustrate the power of machine learning, MICROMINE has built an algorithm to determine and map the spatial extents of core imagery in a core tray photo. The application of this technology will result in the reduction of man-hours required to manually review and analyse core tray photography, the company said.

MICROMINE is incorporating machine learning into its solutions, with the results of the research project leading to the functionality being incorporated into the Geobank data management solution, enabling core tray images to be transferred into the database and displayed in Geobank drill-hole trace along with other downhole data, the company said.

MICROMINE’s presentation is part of APCOM’s technical program, which is presented within six streams: Geostatistics and Resource Estimation; Mine Planning; Scheduling and Dispatch; Mine Operation in Digital Transformation; Emerging Technologies and Robotics in Mining; and Synergies from Other Industries.

Whitehouse will be joined by around 100 international presenters from science and industry at the three-day APCOM conference (June 4-6).

You can read more about the event here.

International Mining is a media partner for APCOM 2019.