How much gold remains to be mined on Earth? How about the lesser-known element, indium, necessary for computer and smartphone displays? With known sources of some essential metals facing depletion within the next few decades, there is more pressure on pursuing alternatives to existing mining exploration technologies. How can we more easily find needed deposits on earth and in space?

Figure 1. Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). (Image courtesy of NASA)

Multispectral imaging is already used in satellite, drone, and aircraft-based systems in the hunt for new mineral deposits on Earth, and it's also starting to be used in space. The next logical step is likely hyperspectral imaging. Routine use of existing hyperspectral systems by the minerals industry has been hampered by the unavailability of systems for industrial use, the high cost of hyperspectral data (when available) compared to typical multispectral data, and the need for additional research into the processing of hyperspectral data.

After a few rough years in terms of growth and investment from 2010 – 2015, with falling prices and low investment, 2017 is starting to look better, and companies may be looking to invest in new technologies to find and extract the resources the world's economy is demanding.

A History of Innovation

Remote imaging has been used in exploration for a long time. At its most basic level, geologists and prospectors would simply point their cameras out of aircraft (first balloons, then airplanes, now drones) to take pictures of the ground below, gathering information on topography and soil that might reveal clues about the location of minerals. With the help of some complicated math, detailed maps could be derived from these photographs. Today, remote sensing has become one of the most important methods for quickly and directly acquiring information about the Earth's surface.

But, visible spectrum photography alone has limitations — weather, daylight, and the simple fact that much of what they were looking for was hidden underground. Much had to be inferred. It wasn't until post-World War II, that new sensing modalities came into play as technologies developed during the war were then applied to commercial applications. Infrared cameras could penetrate inclement weather conditions better than conventional photography and more easily identify the mineral content of soil. Magnetometers could sense disturbances in the Earth's magnetic field to pinpoint metallic ore deposits deep underground. Gravimeters measured the pull of Earth's gravity, which varies slightly based on the position of underground mineral deposits. Radar has many of the same advantages as infrared film, but can also see through limited vegetation, revealing the geological features underneath, day or night, including surface texture and moisture content.

Figure 2. A 2014 high-tech mineral-mapping effort in Afghanistan by the U.S. Geological Survey. (Image courtesy Department of Defense).

Moving imaging systems from aircraft to satellites yielded entirely new possibilities. Government support was critical in initiating this technology. Current technologies include the Landsat thematic mapper and the enhanced thematic mapper (TM) multispectral imager by the United States, and high-resolution panchromatic imaging technology (SPOT) developed by the French Space Agency.

Enter Hyperspectral Imaging

Imaging in the visual range simply didn't offer the detail and information that the mining industry needed. Enter hyperspectral imaging, courtesy of NASA. The technology was first developed in the late 1970s by Jet Propulsion Laboratory, enabling NASA to put hyperspectral imaging equipment in satellites sent to Jupiter and Saturn. While a few private companies built their own hyperspectral cameras, it really took off when NASA made the technology available to researchers and entrepreneurs, even offering grants to test the real-world effectiveness of the technology, including one to Yellowstone Ecosystem Studies.

Hyperspectral imaging makes use of the fact that all objects possess a unique spectral fingerprint based on the wavelengths of visible and invisible light that they absorb and reflect. This reveals a great many details not available in the visible spectrum, such as the difference between greens that represent natural plants, and those that are not natural, such as plastics and metals. Out beyond the visible spectrum, green plastic has different reflectant properties than natural vegetation — even fallen or cut branches have a different fingerprint than growing vegetation.

The commercial impact will be enormous, particularly in the field of mineral exploration. While gold occurs in amounts too small to detect with any available technology, more common minerals like kaolinite and arsenic, which are products of some of the same geological processes, are clearly visible in open landscapes, for example, in much of the American West or the Australian desert. For the diamond industry, kimberlite pipes, the volcanic formations that brought diamonds to the surface, are easy to identify from the air with hyperspectral imaging.

Figure 3. Scientists collected hyperspectral imagery of the Lamar Valley in Yellowstone National Park between 1999 and 2014. By performing a series of classifications on these images and carefully aligning them, we can see how the landscape has changed. (Image courtesy of Yellowstone Ecological Research Center)

The data demands are heavy, putting a lot of pressure on each organization's computing power. For example, when NASA used hyperspectral imaging to investigate the 16-square-mile Superfund site in Leadville, Colorado, it saved years of on-the-ground work, with 45 seconds of satellite imaging. Still, it took another 10 months to crunch the numbers. Only with the development of more powerful computers has this kind of imaging been brought out of the lab. The rapid growth in computing power and storage capacity in widely-available computing platforms has allowed the very large data sets, terabytes and exabytes, of information from airborne instruments to be handled in the time frame required by mineral exploration operations.

Exploring Alaska

The goal of the Alaska mineral resource investigation project is to define the geologic footprint of select deposits using imaging spectroscopy, and regionally extrapolate this knowledge to areas not well characterized. It is expected that the synthesis of results from this multi-disciplinary project will enhance our understanding of the regional geology and be used to develop a predictive exploration model for the identification of base and precious metal-bearing deposits in Alaska and similar remote regions of the world.

Figure 4. Hyperspectral remote sensing data and a multi-proxy investigation for characterizing mineral resources in Alaska.

Looking Deeper and Looking Forward

This new data demands new understanding of the movement of fluids through the Earth. Enhanced hydrologic models will be critical for future mineral exploration. This is also relevant to the effective closing of mines that have completed their life cycle. Models for ore deposits that, when mined, will have minimal impacts on the environment (such as deposits with no acid-generating capacity) and for deposits that may be available to innovative in-situ extraction, will be important for the future.

Technology has progressed to a point where it is now possible to predict three-dimensional images from two-dimensional analyses in some mineral systems. This could ultimately lead to defining liberation in an ore, thus eliminating overgrinding and reducing both energy usage and excessive loss of fine-grained particles.

Currently, a number of research challenges are being addressed for hyperspectral technology, especially for spaceborne systems. These include the development of focal planes with adequate signal-to-noise spectral resolution to resolve mineral species of importance and the capability of acquiring data at a 10-meter spatial resolution while maintaining a minimum swath width of 10 kilometers. The focal planes must also be compact, lightweight, have accurate pointing capabilities, and be robust enough to maintain calibration for long-duration spaceflights.

Where We Stand Now

Routine use of existing hyperspectral systems by the minerals industry has been hampered by the unavailability of systems for industrial use, the high cost of hyperspectral data (when available) compared to typical multispectral data, and the need for additional research into the processing of hyperspectral data. There is still a lot of work to be done. In many countries, national governments are funding, and companies are investing in, system development and deployment as well as basic research on the analysis of hyperspectral data that would ensure these new technologies will be useful for the mineral exploration industry, as well as for a wide range of other users, including land-use planners and environmental scientists.

This article was written by Inder Kohli, Senior Product Manager, Infrared Imaging, Teledyne DALSA (Waterloo, Canada). For more information, Click Here.