However, as in mining, mineral processing, logistics and infrastructure, big data analysis may yet be a key to unlocking significant new value in the industry.
In the final report in a six-part series of technologies set to change mining in future, RFC Ambrian says geophysics, geochemical prospecting, hydrochemistry surveys and new-generation down-hole surveying and core analysis technologies are all moving forward to varying degrees, and could make bigger contributions as the industry grapples increasingly with blind deposits hidden under deep, barren cover.
"We find that much of today's innovation is largely incremental and the difference between new technology and incremental improvement is sometimes blurred," RFC Ambrian analysts, the report authors, say.
They say the application of digitalisation is taking place on the data processing side of exploration, including the use of big data to better understand geologies and predict potential economic resources.
"At the same time there is the potential for the data to be fully integrated into the mine planning, development, and operation as part of the mine of the future," the authors say.
"Once robust data collection has been achieved, the analysis and evaluation of geotechnical data remains fundamental to the engineering design process. Increases in drilling, sampling, geochemical and geophysical methods have produced large volumes of data.
"Increased data volumes have benefited from increased computer processing power with relational databases for systematic storage and programs for data processing and modelling."
A geoscientist can process and visualise huge volumes of data in three dimensions, which is becoming an increasingly valuable way to understand spatial data relationships. Meanwhile, advances in mathematical algorithms, geostatistics, simulations and other state-of-the-art applications are supporting better geoscientific modelling.
RFC Ambrian's report says the increasing complexity of geological data stored in multiple software platforms is driving moves toward automated classification techniques, or machine learning, to interrogate data more deeply than is generally possible using manual methods.
"The routine acquisition of large, multi-band and multi-element geoscientific data sets lends itself to this type of analysis," it says, while cautioning there are "significant challenges" with practical application of machine learning in regional exploration contexts, including non-uniform data coverage, varying data types, model resolution variations, paucity and variability of training points, and signatures that manifest at different scales of observation.
"All these factors require specific attention when attempting to construct a representative prospectivity model using many of the common machine-learning algorithms available," RFC Ambrian says.