AI technology is used for geological mapping

March 2, 2021

AI mapping, a digitalized solution for geological mapping and assessment, is now being tested in the Kiruna mine. At the touch of a button on the telephone, the mine geologist can access real-time data that can be decisive for production. This is a giant leap forward in terms of technology and development. 

AI mapping is much more than just a great leap forward where technology is concerned. The solution also means big gains in terms of environment and efficiency.

“Production geologists in the mine map up to 15 working faces per shift, which often means travelling tens of kilometres by car to reach all of the production areas. Thanks to AI mapping, we can do this work much more efficiently and effectively,” says Jonathan Sumelius, geologist at LKAB in Kiruna.

But let’s start from the beginning. In the mine, the production geologists map the production areas. They are the link between mine development crews and production, since the development areas must be surveyed, or mapped, before production crews can be given the go-ahead to proceed with mining. In other words, the mapping data are decisive for production.

“When we do mapping, among other things, we look at rock types, ore quality or grade, and contacts between ore and different waste rocks. We also look at structures and estimate the strength of the rock mass,” says Jonathan Sumelius.

The work of the production geologist, in the conventional sense, involves a very close relationship with the rock. Here, the naked eye is the most important tool.  To experience, see, hear and feel the rock is essential for being able to make a reliable assessment of the status and condition of the rock.

“The Kiruna orebody is a single, continuous slice of magnetite, which means that it is not especially complex. We frequently provide reliable forecasts and can usually foresee when we are going to hit ore or waste rock during drifting. This is one of the reasons why it is appropriate to apply AI mapping in the Kiruna mine,” says Jonathan Sumelius. 

AI-gavelkartering, AI face mapping, is the name of the app that is used to supply the image processing program COCO Annotator with data. By scanning a QR code in the app, with which mine development machines are also linked, the program can access information as to exact locations in the mine. Then, photographs are taken and the images are processed directly in the program in real time. The AI technology is revolutionary, since it enables the program to improve all the time; the more data and images it acquires, the better it gets.

“We have been developing AI mapping for a little over a year, but much remains to be done. We need about a thousand images before we can be sure of a reliable result,” says Jonathan Sumelius.

When an image is processed in the app the program uses an algorithm that is continuously refined and improved. The result is based on three central parameters: iron ore, barren rock and cracks, and the visualization shows a clear indication of occurrences of waste rock.

“It is really exciting to work with this and to take part in the development of the technology. It’s a whole new way of thinking and working,” says Jonathan Sumelius.

AI mapping technology also enables mapping of a larger area of the mine, and doing it faster, but with the same number of personnel. This makes it possible for the mine development crews, who work in the mapping areas, to use the app to deliver data to the geologists in real time. This is a big step forward, not least for the environment.

Jonathan Sumelius.jpg

Jonathan Sumelius, geologist at LKAB in Kiruna.