FEATURE: HYPERSPECTRAL
Democratising applications Innovation in the field stems from the interplay between government, academia, and industry. Academics bring theoretical advancements and novel sensors and algorithms, building the foundational knowledge on which every advanced industry is built. Governments boost R&D spending and offer infrastructure, while companies make the whole process more efficient and try to address user needs in a feasible and economical way. The good news about the last part is that there is no shortage of user needs. “One of the most crucial applications in
the short term is soil health monitoring,” says Zevenbergen. “New initiatives from environmental and agricultural agencies around the globe aim to accurately map, monitor and optimise soil health. But this currently relies heavily on in-situ data collection and scaling regionally is impossible. “The leading solutions are with space- based SWIR data; however, researchers are demanding more bands in the >2000nm range and more frequently collected imagery. As the barriers to build and launch hyperspectral sensor payloads decrease, soil health analysis applications should skyrocket in the coming years.” Meanwhile, Doumet mentions mining
applications. He is already working with companies who want to explore large
areas for signs of mineral resources: “We’re using hyperspectral imagery to find underground deposits or at least some indications of mineral characteristics.” There’s also agriculture. Hyperspectral imaging can find applications in species identification, monitoring crop health, yield estimation, and even early stress detection. Similar approaches can help to manage ecosystems, from forests to delicate coastal environments. Water quality and invasive species could be detected using hyperspectral imaging, and “new applications are being discovered at an astonishing rate”, says Goodman. These are all applications that can be used by small companies, perhaps even by individuals. For society as a whole, large-scale
applications such as monitoring climate change or tracking methane leaks (both of which are excellent applications for hyperspectral Earth observation) are important. But for a farmer, whether or not they should worry about soil quality or invasive species is much more important. For a surveyor, having access to information on where to focus ground- based efforts can make the difference between success and failure. For all these types of users, having access to localised, precise information is what matters. Doumet says this is what we’re heading towards – and sooner rather than later.
14 IMAGING AND MACHINE VISION EUROPE DECEMBER 2023/JANUARY 2024
“Hyperspectral imaging is going to significantly impact the satellite industry”
A watershed moment “Hyperspectral imaging is going to significantly impact the satellite industry, not just as a technological advancement but as a catalyst for broader network effects,” says Zevenbergen. He echoes a feeling that seems to be
shared among the entire industry – that the increasing demand for hyperspectral imaging will first spur more innovations in satellite technology. Whether it’s more sophisticated sensors or better satellite designs that include onboard processing capability, this will further push the boundaries. Then, with improved data processing
techniques (largely fuelled by machine learning algorithms), feature extraction and classification from large datasets will become easier, enabling more users to gather value from hyperspectral data – bringing more demand into the field and further promoting innovation. “The future of hyperspectral is now,” concludes Goodman. I
@imveurope |
www.imveurope.com
Shutterstock/VadimSadovski
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32
