Finding gold is a notoriously tricky business, as generations of prospectors have discovered. The ore is usually buried deep underground, so the best way to locate a literal goldmine is to look for indicator minerals that are known to be associated with gold’s presence: Copper, lead, zinc, cobalt, and others. Often this process is no more sophisticated than a person with a hammer and a hand lens, chipping away at rocky outcrops, searching for mineralogical clues.
But now there’s another way. “Whether you’re talking about gold or diamonds or any other valuable metals or minerals, we can detect the indicator materials from space, which then narrows down the areas that mining companies need to explore,” says Awais Ahmed, founder and CEO of Pixxel, an Indo-US startup that is commercializing the science of hyperspectral imaging (HSI) via satellites in low Earth orbit.
Hyperspectral cameras capture three-dimensional image datasets containing not only spatial information about a scene but also spectral information. They can effectively “see” and record images across hundreds of bands of the electromagnetic spectrum, including regions in both the ultraviolet and infrared, as opposed to the very limited number of wavelengths that are visible to regular cameras or the human eye. As Pixxel’s cameras sweep over the planet, analysis of the data they capture provides detailed information on the physical and chemical properties of surface rock formations, thanks to the unique way every substance reflects light. In other words, if an optical camera merely records what something looks like, a hyperspectral camera can also tell you what that thing is made from.
Pixxel currently has three satellites orbiting the globe, with six scheduled to follow in 2024, and 18 the year after. The plan, ultimately, is to bring global coverage updated every 24 hours. The use of HSI in Earth orbit was until recently largely the domain of governments and scientists. The emergence of Pixxel and a handful of other companies vying for a share of this new commercial market is the result of two converging trends. One is the growth of private satellite launch companies, which have opened up cost-effective access to Earth’s orbit for a large number of new, innovative startups. The other is recent advances in HSI technology that have brought increased image resolution while also reducing the size of hyperspectral cameras, making them an easy fit for constellations of small satellites. One study suggests that the HSI systems market is projected to grow from $847m in 2024 to $1.5 billion by 2029.
Pixxel’s technology can analyze any object on the planet’s surface, and Ahmed speaks enthusiastically of the many use cases. These include detecting changes in the health and status of agricultural crops; monitoring the carbon footprint of mining operations by tracking the waste and poisonous materials in their dumps; spotting pollution from pipeline leakages; and assessing the feasibility of construction projects.
Pixxel’s customers currently come mainly from the agricultural, oil and gas, and mining sectors, plus some governments, and they are offered access to the company’s data platform, which allows user-friendly insights to be generated from the information gathered by its satellites. “Hyperspectral data is complex,” says Ahmed. “The platform removes the need for our customers to hire specialists.”
In the future, Ahmed dreams of working with space agencies to map the moon, Mars, and asteroids. There may be valuable material that could be mined for use in-situ to aid space exploration, or brought back to Earth for use terrestrially. But for now, he’s mainly on a mission of education, explaining the benefits of satellite-based HSI to potential customers. “Originally we just talked about it,” he says. “But now we educate them by showing them what hyperspectral image analysis from our data can actually do.”
Down on Earth, there is an emerging trend for bringing accurate and detailed—and often real-time—HSI into specialized applications. These range from examining living tissue during surgery to monitoring food quality on production lines.
At the Hyperspectral Imaging Centre at the University of Strathclyde in Glasgow, Co-Director Dr. Paul Murray is currently researching the use of HSI in the nuclear industry, along with his colleagues Professor Stephen Marshall and Professor Malcolm Joyce from Lancaster University. One project involves developing hyperspectral technology for the autonomous inspection and monitoring of nuclear fuel pellets during production in the UK. The idea is that HSI can find tiny flaws or cracks in the pellets, allowing a responsive manufacturing system to detect evolving flaws in the machinery and thus nip problems in the bud.
A second project is looking at new ways to use HSI to map environments where nuclear fuel debris may be present. One of its first applications would be at the Fukushima nuclear power plant in Japan, where a major earthquake followed by a tsunami in 2011 resulted in core meltdowns in the reactors. Removing the melted fuel remains a significant technical challenge, and identifying it in the first place is a major part of that. A team of engineers and scientists from the UK and Japan is exploring the use of HSI together with other sensor technologies, signal processing and data fusion to do just that. “We anticipate that the research will lead to new and highly valuable inspection technology which can support nuclear decommissioning in Japan, the UK and around the world,” says Murray.
Another major trend in the HSI field is the quest to create easily portable cameras that can both capture and process hyperspectral data on the spot. The Holy Grail would be something like Star Trek’s famous Tricorder, the handheld device used in the sci-fi franchise to instantly provide detailed readings on the environment of alien planets. So far, the smallest portable hyperspectral camera on the market that can also analyze data is the size of a large SLR camera, weighs 1.3kg, and has a similarly weighty price tag. A potentially much cheaper alternative, which is attracting interest from some researchers, would use AI to “construct” hyperspectral images from clues in RGB images taken by a standard camera. However, the proposed solutions require a considerable amount of AI training, which can be constrained by the relative scarcity of HSI datasets.
“In the next five to ten years I hope that the technology will become smaller, lighter, and much more affordable,” says Murray. “I’d be surprised if we had hyperspectral imaging on our smartphones soon, but I’d like to see far more handheld, portable devices that give you an image instantly on the screen and allow you to identify all the materials in the environment.” There would be many possible use cases, he suggests, from a farmer mounting a device on their tractor to check their crops as they drive around, to an emergency worker instantly identifying hazardous material in an industrial accident.
Ultimately, Murray says, he would like HSI to become so accessible that it’s the norm. “Imagine if anyone who’s currently using a camera for any commercial purpose, whether it’s inspection of oil and gas infrastructure or security or whatever, was able to say, ‘Why don’t we just put a hyperspectral camera there?’” he says. “In twenty years’ time, that could be the case.”
David Rosewell | Group Director for Science and Technology at QinetiQ
When dealing with critical missions out in the field, technology that increases available information about the environment, such as hyperspectral imaging (HSI), will always provide an operational advantage. The anticipated cost and size reduction of HSI sensors will increase their proliferation in defense uses for deployments across air, land, sea, and space.
HSI can help identify specific activities and targets, based on materials and composition. However, more information also introduces the problem of “drinking from a waterfall”; it requires improved means of sorting wheat from chaff, getting to the nugget of greatest importance and disseminating it to those who need it most urgently, as quickly as possible.
We expect armed forces with advanced and widely used HSI sensors to have information dominance over adversaries, but they will also need advanced data analytics, machine learning, and likely artificial intelligence if they are to be more effective.
- Mechanical human augmentation. Whether it’s additional limbs or smart exoskeletons, machinery is helping humans upgrade their natural capabilities.
- Power beaming. Sending power wirelessly over long distances could transform everything from electric vehicles to offshore wind farms.
- Biohybrid robots. Combining artificial and organic parts, biohybrid robots offer advantages such as self-repair and agility.
- Neuromorphic computing. Inspired by the brain, neuromorphic chips aim to equal the speed, efficiency, and intelligence of the human mind.
- Gene-editing and enhancement. Advances in biotech are spurring scientists to explore how genomes can be tweaked to make ecosystems more sustainable.
To find out more about QinetiQ, click here
To find out more about WIRED Consulting, click here
