Headwall Photonics Blog

Hyperspectral Imaging Technology a New Frontier for KAUST in Saudi Arabia

Posted by Christopher Van Veen on Thu, Oct 08, 2015

King Abdullah University of Science and Technology (KAUST) is a public research institution in Saudi Arabia. By any measure it is a very young university, founded in 2009. But in that short span of time it has rapidly grown to accumulate an astounding number of research and citation records.

Being a science and technology university, KAUST focuses on traditional subjects such as math, electrical engineering, and computer science. KAUST is already well versed in spectroscopy, having a burgeoning lab full of instruments that can peer into the chemical underpinnings of minerals, plants, and crops. The lab is outfitted with spectroscopy, chromatography, and mass spectometry instruments tasked with learning more about trace metals analysis, wet chemistry, and surface analysis.

But one area of study is on pace to become its most popular: Earth and Environmental Sciences. Here, students and faculty pore over ways to use new technology in learning more about precision agriculture, water resources, and atmospheric conditions. This takes spectroscopy out of the lab and into the air, specifically using drones and UAVs.

KAUST_Nano_1_with_caption_small.jpg

Spectroscopy takes several different forms, but multispectral and hyperspectral are of primary interest within the scientific research community. The primary difference between the two is that hyperspectral imaging technology provides complete spectral information for every pixel in the scene...literally hundreds of bands. Comparatively, multispectral will only detect a handful of bands. In other words, it can mean the difference between discovering an invasive disease in a crop field or vineyard and missing it altogether. The Headwall sensors have a wide field of view (FOV) coupled with aberration-corrected optics, meaning that image data is as crisp and precise along the edges of the FOV as it is directly underneath the flight path. More data is collected for each mission, making the data-collection project more efficient. With battery life being key, efficiency matters.  Armed with spectral libraries that define the chemical composition of everything the sensor ‘sees,’ scientists have the ability to look well beyond what is actually ‘visible.’ 

Matthew McCabe, a professor at KAUST, recognizes the value of real-time environmental analysis in the work his Hydrology and Land Observation (HALO) Group does. “We are interested in exploring hyperspectral sensing to better understand plant health and function, particularly as relates to agricultural settings,” said McCabe. “The capacity to retrieve information on plant health is of interest not just for the obvious and important monitoring of crop state, but also in better constraining coupled water-energy-carbon models of vegetation systems.” According to McCabe, these models are generally poorly constrained so a system that enables for investigating plant health and condition in near real-time is of much interest. 

Already familiar with less precise multispectral instruments, McCabe is now investing in airborne hyperspectral sensors that see more and can deliver vast amounts of spectral data. “While we already employ multispectral sensors covering bands known to inform upon plant systems, it is the capacity to further expand our knowledge of plant spectral response that we are most interested in,” said McCabe. “Determining new spectral relationships in plant behavior and response is an area of research that hyperspectral imaging can really drive.”

KAUST_Nano_2_with_caption_small.jpg

The field of crop science is a key deployment for hyperspectral imaging. Here, McCabe chose Headwall’s Nano-Hyperspec sensor that covers the core Visible/Near-Infrared (VNIR) range from 400nm to 1000nm. Most anything that a crop scientist wants to ‘see’ will be found in that VNIR range. “We wanted a very precise spectral imager that operated in this important VNIR range,” said McCabe. “We needed to ‘see’ spectral information for every pixel within the field of view, and we knew that Headwall’s Nano-Hyperspec had very precise edge-to-edge imaging performance that would optimize the flight efficiency of our UAV.

With the newly acquired Nano-Hyperspec mounted aboard a KAUST-engineered quadcopter, McCabe and his team will set off to learn about plant-specific spectral-traits that provide direct insight into health and productivity functions. “This is one of our research goals,” said McCabe. “Hyperspectral sensing and analysis allows us to explore this exciting area of research in ways that multispectral cannot.”

Choosing a sensor and deploying it properly on a UAV is a challenge, especially since the technology is still relatively young. “We needed a sensor package that matched our UAV,” said McCabe. That meant it had to be lightweight and small, and it had to have integrated data storage for the many gigabytes of spectral data pouring into the sensor while aloft. The Headwall solution added a GPS/IMU and full software control for an overall package that put McCabe in the air far sooner than he’d otherwise expect. “Partnering with Headwall gave us months of headway in terms of getting in the air and collecting great data,” said McCabe.

And why Headwall? “We knew the company was a strong and well respected brand in the spectral sensing domain, with a long history of product excellence,” said McCabe. “We had researched a number of competing solutions, but Headwall turned out to be the most professional and competent among them.” McCabe also noted that numerous colleagues and scientists were currently using Headwall hyperspectral instruments with great success.

Moving forward, what are some of the exciting plans in the area of environmental analysis using Headwall’s hyperspectral system? “Our initial research will be focused on the indirect retrieval of plant pigments such as chlorophyll and carotenoids,” said McCabe. This spectral data will provide information on water use and stress condition of agricultural systems. “Ultimately, we are interested in better understanding plant water use through transpiration, but there are also opportunities for better constraining crop-yield through routine spatial sampling that is available when coupled to a UAV,” noted McCabe. “Hyperspectral sensing allows for a number of innovative ways to explore these ideas. But there are also clear opportunities beyond crop science. Indeed, we see applications across many of the multi-disciplinary research areas we are engaged in.” Interestingly but not surprisingly, McCabe is thinking well beyond UAVs: “We also have active projects in satellite validation and expect to engage with collaborators on both small and larger scale crop stress monitoring and even disease mapping using sensors mounted aboard L.E.O. commercial satellites.”

Of course, Headwall will be there when the time comes.

Click to edit your new post...

Tags: Remote Sensing, UAS