New space collaboration eyes hyperspectral imaging

A Europe-Australia collaboration aims to explore the use of hyperspectral imaging in space and find ways to reduce the bandwidth needed to send images to Earth.

It solves a problem many of us have – it’s one thing to have hundreds of high-resolution, information-rich satellite images delivered daily to your inbox. It’s entirely another to make full use of them.

And that’s the subject of the new international agreement between Australia’s space research centre Smartsat Cooperative Research Centre (CRC), and the European Space Agency’s Φ-lab (Phi-lab).

“We have also identified new collaboration opportunities between researchers at the University of Queensland and Φ-lab to develop advanced AI capabilities using hyperspectral satellite imagery,” SmartSat CEO Professor Andy Koronios said shortly after signing the agreement in Rome. “This will create the next generation of predictive intelligence that can forecast and monitor agricultural and environmental activities and disaster events with greater accuracy.”

Hyperspectral sensors add a multitude of additional layers to traditional space-based photography. Every pixel can include hundreds of wavelengths, ranging from the infrared to the ultraviolet and beyond. This creates a spectrum that can “fingerprint” any object it captures.

The uses for such capabilities are only just beginning to be understood.

But its potential has already been recognised for weather forecasting, water quality tracking, land-use monitoring, crop and land health analysis, forestry management, pollutant detection and disaster assessment.

Meet Kanyini

South Australia’s technology demonstrator 6U CubeSat Kanyini will carry the HyperScout multispectral imager when it is launched later this year. This camera, capable of capturing infrared and visible wavelengths, was developed by the European-based Cosine Measurement Systems in conjunction with the ESA’s Φ-lab. 

This camera aims to explore ways to sift and pre-select relevant imagery before the bandwidth-intensive files are downloaded to Earth. This will involve new low-power “edge computing” (space-based processing) technology.

Koronios says the ESA’s earlier experience with hyperspectral cameras and AI analysis through its 2020 Φ-sat-1 satellite mission offers “significant collaboration opportunities”.

This will include an exchange program for research personnel and internships for students specialising in onboard AI and synthetic aperture radar systems.

“Australian Space Agency chief Enrico Palermo welcomed the collaboration.

“Earth observation is one of the strongest examples of how space technologies improve life on Earth – from forecasting the weather to responding to natural disasters and mitigating climate change. This agreement will help bring together bright minds in Australia and Europe to develop space solutions integrating AI technologies and research.”