Emerging solutions to beat the hyperspectral threat

If your job is growing crops, then the current global surge in hyperspectral sensor systems is great news. By allowing the different chemical constituents of a piece of terrain to be quickly and easily identified, such systems can help farmers to establish which parts of their fields need more water, fertilizer or pesticide.          
But if you’re a soldier in a combat zone trying to avoid detection by the enemy, hyperspectral systems represent a major threat. Their ability to determine whether small segments of a scanned surface are of natural or synthetic origin means that they can potentially identify deployed camouflage solutions – putting troops and mission success at risk.

Hyperspectral sensors capture reflected radiation from hundreds of spectral bands.

While the widespread use of hyperspectral sensor solutions is in its infancy, numbers of both satellite- and drone-mounted systems are surging rapidly. Finnish company Kuva Space, for example, has ambitious plans to launch 100 hyperspectral mini-satellites and to sell the data it collects commercially.

It is time to face the growing satellite threat

It’s not if they’re looking. It’s if you let them see you.

Why signature analysis is crucial for combat survival

 Commercial drone-mounted hyperspectral cameras from countries such as China are also now available for a range of applications, from minerology to greenhouse gas monitoring. This sudden explosion in hyperspectral technology has left both camouflage manufacturers and global armed forces reappraising their approaches to deceiving the enemy.

So how do hyperspectral sensors work? What do they mean for camouflage systems? And what techniques and technologies can potentially be used to defeat them?

Hundreds of spectral bands 

Today, a wide range of active and passive sensors are used in the defence sector to identify deployed military assets. Among the most common tools are multispectral sensors, which capture reflected radiation within five to 10 bands of the electromagnetic spectrum. Users of multispectral sensors are able to identify military assets by recognising their distinctive geometric shape and spectral contrast to the surrounding area. High-quality multispectral camouflage solutions such as those manufactured by Saab’s Barracuda business unit are able to defeat such sensors by hiding the true signatures of assets and mimicking the signatures of the surrounding environment.

Hyperspectral sensors, by contrast, capture reflected radiation from hundreds of spectral bands , meaning they are far more sensitive. This means, with the help of algorithms and AI, users can identify and quantify specific materials within a scanned area. While a multispectral sensor is likely to be fooled by a camouflage system whose colour, reflective properties and thermal properties matches its surrounds, a hyperspectral may be used to identify the presence of plastics in the camouflage. As plastic is not commonly found in natural environments, the end users may be able to deduce that camouflage is being used.

While information on hyperspectral sensor use in the defence space is often tightly guarded, a range of satellite and drone solutions are known to be in use. For example, the Elbit Systems Hermes 900 Kochav is an Israeli-manufactured long-endurance UAV with hyperspectral capability. It is in use in a range of countries globally. 

With military use of such equipment growing rapidly, armed forces need to find ways to defeat such sensors to preserve the tactical advantage of concealment and deceit. Thankfully, there are current and emerging technologies and techniques that have the potential to rebalance the equation.

Back to nature

One approach that is effective right now is to employ high-quality multispectral camouflage and cover it with fresh vegetation from the surrounding area. Doing so covers the plastic materials used in the camouflage system with non-synthetic material. If scanned by a hyperspectral scanner, the area will appear more or less the same as its surrounds. Meanwhile, the radar attenuating capabilities of the camouflage system below remain unchanged, protecting concealed assets against radar detection. One downside of this approach is that the vegetation needs to be changed daily, as dead plants have a different spectral fingerprint to living ones.
This approach can be optimised through the use of decoy camouflage systems, where poor quality systems easily detected by sensors are set up to draw the enemy’s attention away from the real assets. 

Three interesting avenues of research

In the longer term there are a number of promising technological solutions which are being explored.

One is to replace the polymers used in current camouflage solutions with plant-based alternatives. Pigments, textiles and binders made from forest products and minerals would stand out far less to hyperspectral sensors than their plastic counterparts, potentially allowing deployed troops to avoid detection longer. One downside of this is that non-synthetic materials typically deteriorate far faster than synthetic materials, meaning camouflage systems may need to be replaced more frequently. However, a major advantage may be reducing the camouflage industry’s dependence on polymers supplied by nations in South-East Asia. Sweden, for example, has a highly developed forestry industry and a switch to non-synthetic camouflage could boost in-country manufacturing capabilities in times of conflict.

Another interesting option is to use artificial intelligence to develop synthetic materials that mimic the properties of non-synthetic materials when viewed through a hyperspectral scanner. A molecule that has the same spectral reflection as pine foliage, for example, could be coated onto synthetic camouflage intended for use in pine forests. Because the coating matches the spectral signature of its surrounds, it is unlikely to be identified as an object of interest by sensor solutions.

Yet another promising approach is to change the patterns on conventional synthetic camouflage. There appears to be potential to create designs that confuse or dazzle hyperspectral sensors. This approach might lead to deployed assets being completely ignored or having to spend time and deploy assets to investigate further.

While the emergence of these potential solutions provides hope that hyperspectral solutions can be defeated, dedicated work on these options is required now. The development and roll-out of hyperspectral systems is likely to follow Kurzweil Curve and grow exponentially rather than lineally. This means large numbers of these systems are likely to be in military use within the next five years.