Hyperspectral imaging is finding a home in so many interesting places, among them the fascinating field of cultural preservation. Conservation care professionals across academia and the museum world are tasked with learning as much as they can about the treasures under their care. These treasures range from artifacts such as vases, to paintings, documents, and maps.
In all cases the objective is to non-invasively increase the body of knowledge. Are there features that are invisible by any other means of analysis? Are there chemical pigmentation signatures on paintings that spectroscopy can ‘see?’ Are there any hidden writings that can be uncovered? Hyperspectral imaging can help conservation-care experts determine origins, dates, materials, and other characteristics useful to their work. Indeed, spectroscopy can also help improve the preservation of these treasures by uncovering evidence of similar efforts done years or decades previously. Hyperspectral imagers offer scholars, curators and conservators unique advantages:
- Enhance faded or hidden features-text/signatures
- Detect restorations and repairs via chemical signature
- Monitor and track changes of the object, or repairs and restorations
- Identify local material components for proper repair
- Assess original coloring and pigmentation
On December 9 at The Natural History Museum in London, Headwall organized and sponsored a workshop and seminar on hyperspectral imaging in the collection-care industry. Noted experts from worldwide universities, museums, and libraries came to hear about how hyperspectral imaging can help unlock hidden secrets while advancing the overall body of knowledge of the treasures under their care.
Mr. David Howell of The Bodleian Libraries spoke about building a suite of non-destructive imaging techniques. Mr. Chris Collins of The Natural History Museum spoke about assessing fading in natural history specimens. And Christina Duffy of The British Library discussed their use of multispectral imaging on the treasures under their care (including the Magna Carta!).
Setting the stage for the day was Mr. Kwok Wong, who serves as Headwall’s Senior Systems Applications Engineer. Kwok has done considerable work with The Museum of Fine Arts (MFA) in Boston, imaging a Mayan Vase and other artifacts. Kwok explained the basics behind multispectral and hyperspectral imaging and the kinds of valuable information that can be collected in a non-destructive, non-invasive manner.
Dr. Greg Bearman, a noted expert in the field of spectral/chemical imaging within the collection-care industry, discussed his impressive work to date and how the techniques can best be applied. Dr. Bearman’s examples included paintings, documents, and artifacts...with each requiring a slightly different approach depending on the spectral ranges that need to be covered.
Guests were encouraged to bring samples of their treasures for Headwall to image during the day. A few of the attendees did so, and Headwall had its VNIR (Visible/Near-Infrared) Starter Kit operational in the room. Attendees could see first-hand how the science of spectroscopy can be used to further their preservation and analytical efforts.
Most often, the collection-care industry cares most about imaging in the VNIR (380-1000nm) and SWIR (950-2500; short-wave infra-red) ranges. Imaging in the VNIR and SWIR has a number of important and interesting applications for Cultural Heritage because this type of imaging technology provides a more complete representation of the entire field of view. This is a critical distinction because true context is provided on what are typically heterogeneous objects; by comparison, point sensors can only sample discrete locations. Imaging in the VNIR has been used since the mid 1990s for texts and paintings. For texts, the application is typically content; for example, reading palimpsests and faded or damaged texts and maps. For art, the application is typically color and pigment mapping. SWIR imaging offers the possibility of chemical imaging, allowing the conservator to monitor and track chemistry changes over time.
Since little or no preparation of the document or artifact is necessary, this non-destructive spectral technique is invaluable for a wide range of conservation research relating to changes in color, chemical and substrates. Within the field of view of the Hyperspec® sensor, hyperspectral imaging provides quantitative spectral information for all wavelengths across the complete spectral range of the sensor.
The key to spectral data is calibration; well-calibrated datasets can be compared and analyzed over time and between multiple users. There is an existing and significant body of spectral analysis, classification and mapping algorithms and software available to work with spectral data. Most of this software has been developed over the last 20 years for satellite remote sensing and is easily available.
The job of the hyperspectral sensor is to collect image data and then assemble this valuable information into a ‘datacube,’ which represents a data set that includes all of the spatial and spectral information within the field of view.
Niki Tsongas, Congresswoman from the 3rd Congressional District in Massachusetts, visited Headwall today to meet with Company officials and speak to employees. Ms. Tsongas applauded Headwall’s focus on technical leadership across its core markets. “I’m fascinated by all the exciting applications for your products,” Tsongas noted. “My work in Congress is aimed at strengthening the entrepreneurial spirit I see when I visit companies like Headwall.”
During the visit, Congresswoman Tsongas toured Headwall’s Fitchburg facility and saw firsthand how the Company’s vertically-integrated approach moves spectral imaging sensors from design to production very rapidly. The sensors, used by industry and government, collect a complete ‘spectral picture’ of whatever is within the field of view. This can be from a satellite, a manned aircraft, a small UAV, or along a high-speed inspection line where product quality can be determined by hyperspectral imaging.
During the ‘town hall’ meeting with employees, Tsongas fielded questions from employees on a range of topics, including her position on STEM education (science, technology, engineering, and mathematics). “Education in these areas represents the catalyst for companies like Headwall to flourish,” noted Tsongas. “You need people who can hit the ground running, and education is fundamental to achieving a labor force that is ready to go in very challenging areas across science and technology.”
Headwall CEO David Bannon thanked Congresswoman Tsongas for visiting Headwall. “We’re very honored to have you here today because it reinforces Washington’s support for small, entrepreneurial, technology-driven companies like ours.”
Congresswoman Tsongas was elected to the United State House of Representatives in a 2007 special election, becoming the first woman in 25 years to serve in Congress from the Commonwealth of Massachusetts. She represents the Massachusetts Third District, which had previously been known as the Fifth District until her most recent reelection in 2012. Tsongas holds the same seat that was held three decades earlier by her late husband, former Congressman, U.S. Senator and presidential candidate Paul Tsongas. The Third District spans portions of Essex, Middlesex and Worcester counties.
Tsongas serves on the House Armed Services Committee, a position she sought out when first elected. In 2013, Tsongas’ hard work led to her being named to a leadership position as the top Democrat on the Subcommittee for Oversight and Investigations. The Third District has a long history of military service, which is reflected both in the number of residents who serve in the active duty military as well as in the numerous veterans who call the Third District home. Tsongas also represents one of the largest concentrations of defense related employers in the country that manufacture the products, develop the technology and create the jobs that keep our nation strong and our service members safe.
As a member of the Armed Services Committee, Tsongas has pushed for development of lightweight body armor and new measures to better prevent and respond to incidents of sexual assault in the military.
Tsongas also serves on the Natural Resources Committee, which oversees legislation related to domestic energy production, National Parks, rivers, forests, oceans and wilderness areas.
More can be learned about Congresswoman Tsongas at her official web site.
High-performance imaging sensors on small, commercial UAS will assess ocean and sea ice variability in Arctic zones
FITCHBURG, MA - OCTOBER 9, 2014: Headwall Photonics has delivered two high-performance hyperspectral imaging sensors to Columbia University as part of its Air-Sea-Ice Physics and Biogeochemistry Experiment (ASIPBEX). ASIPBEX is part of a larger international collaborative investigation of Climate Cryosphere Interaction with colleagues from Spain, Germany and Norway. This crucial remote-sensing project will use a high-endurance unmanned aircraft system (UAS) to investigate climatological changes present in the Arctic Ocean around Svalbard, Norway. The instrument payload comprises two Micro-Hyperpsec sensors; one will cover the Visible-Near-Infrared (VNIR) range of 400-1000nm while the other will cover the Near-Infrared (NIR) range of 900-1700nm. Together, the sensors will be crucial in detecting indicators of sea ice physics, solar warming and global carbon cycles.
"We chose the Headwall sensors for several reasons," stated Christopher Zappa, a Lamont Research Professor at Columbia's Lamont-Doherty Earth Observatory. "The very high resolution allows us to collect and process vast amounts of spectral and spatial data upon which our research and analysis depend." The wide field of view of the Headwall sensor combined with aberration-corrected optics also contributes to overall flight-path efficiency. The UAS allows scientists to measure in places that typically are impossible to get to using ships or manned aircraft. This opens up the possibility for transformative understanding of the climate system. "Since we're using a UAS, we depend on 'seeing' as much of the ocean surface as possible, minimizing any aberrations or unwanted artifacts along the edges of the field of view," noted Prof. Zappa. The combination of Micro-Hyperspec and Headwall's advanced Hyperspec III airborne software allows for the successful collection, classification, and interpretation of the spectral data collected during each flight.
This particular deployment for the ASIPBEX project is fundamental to Headwall's strategy of advancing the science of remote sensing aboard small, commercial unmanned aircraft systems. "Hyperspectral represents a crucial payload for any manned or unmanned deployment," noted Headwall CEO David Bannon. "But significantly notable is that the UAS has become a 'go-to' platform. This means not only smaller and lighter sensors, but also integrated solutions that factor in everything from LiDAR and data-management to post-processing tasks such as ortho-rectification that our software can handle." Because the Micro-Hyperspec sensor uses high efficiency diffraction gratings in a concentric, optical design, imaging performance and signal-to-noise are both maximized. The patented optical design provides a package that is rugged and robust for airborne use in harsh environments such as the Arctic ocean.
The Observatory for Air-Sea Interaction Studies (OASIS)
Led by Professor Christopher Zappa, the Observatory for Air-Sea Interaction Studies (OASIS) conducts research in a variety of fields focused on the oceanic and atmospheric boundary layers. These include wave dynamics and wave breaking, air-sea CO2 gas exchange, non-satellite remote sensing and boundary-layer processes. Affiliated with the Lamont-Doherty Earth Observatory (LDEO) and Columbia University, OASIS is involved in joint projects with the Polar Geophysics Group of LDEO, Yale University, the University of Heidelberg, the University of Connecticut, and the University of New South Wales and participated in various large multi-institution projects such as CBLAST-Low, GasEx, VOCALs, RaDyO, DYNAMO.
The group develops and deploys instruments including infrared, multispectral, and polarimetric cameras on different fixed and mobile platforms such as ships, aircrafts, buoys. The study areas range from laboratory wind-wave tanks, Biosphere2, to local rivers and estuaries, to shelf seas and polynyas, to open ocean from the poles to the equator.
For information contact:
Professor Christopher J. Zappa, Lamont Research Professor
Lamont-Doherty Earth Observatory
At Headwall we've been busy listening to the market. When it comes to airborne remote sensing, the market is telling us that they favor UAVs (unmanned aerial vehicles) of all kinds: fixed-wing, multi-rotor, and so on. There's no end to the number of companies producing UAVs globally. Because many UAVs produced today are very small and affordable they are 'within reach' of those with even modest means. Universities represent one key market where the use of UAVs is rapidly increasing. Full of scientists and research departments, universities around the globe see these small and light UAVs as a perfect platform from which to launch their exploratory studies. They are affordable, easy to assembly and transport, and (especially with multi-rotor models) can take off and land within a very small footprint.
But alongside all this enthusiasm for UAVs, there are many who frown upon these airborne vehicles and see them as a nuisance. Indeed, they can be a nuisance when used for trivial pursuits. In densely-populated areas they certainly can be more than an annoyance...they can be dangerous. But largely, the work we are seeing our customers undertake with hyperspectral imagers attached to UAVs is very valuable work indeed. And it takes place far from the hustle and bustle of any urban landscape. For example, precision agriculture is made more valuable because there are key indices to plant health and physiology that are readily seen from above than from below. Certain disease conditions are ‘visible’ using hyperspectral imaging, especially with high spectral and spatial resolution found on all Headwall sensors. Other research pursuits include environmental analysis, geology, pollution analysis, and so many more. These are very good and valuable scientific efforts made moreso by the UAVs that enable these precision instruments to 'fly.' The marriage between hyperspectral and UAV seems to be a perfect one, especially when you consider how much ground can be covered with one of these flying wizards. And especially when you realize that hyperspectral imaging fundamentally requires that movement needs to occur. In other words, hyperspectral was meant for airborne deployment. Where a Jeep can’t go, a UAV can. And furthermore, more ground can be covered with a UAV, meaning more efficient data collection over rugged and inaccessible landscapes.
As UAVs get smaller and lighter, users run headlong into the issue of payload: UAVs are limited with respect to what they can lift. Whatever else a UAV is asked to carry, it needs to lift batteries. Then comes the instrumentation. Headwall’s Nano-Hyperspec was just introduced for the VNIR (400-1000nm) spectral range. Most (but not all) of the things a research scientist might wish to ‘see’ are visible in this spectral range. But we did a couple things with Nano-Hyperspec that helps the payload issue. First, the size and weight are well below previous sensor offerings. Its size (including lens) is a scant 3” x 3” x 4.72” (76.2mm x 76.2mm x 119.2mm), and its weight is less that 1.5 lb. (0.68kg). Best of all, this includes on-board data storage of 480GB. That’s about 130 minutes at 100fps.
Aside from making Nano-Hyperspec smaller and lighter than other hyperspectral sensors, a key differentiator comes from embedding the data storage within the enclosure while providing multiple attach points for the GPS/INU. Another key attribute is the inclusion of the full airborne version of Headwall’s Hyperspec III software, which includes a polygon flight tool for sensor operation and a real-time Ethernet Waterfall display. While the work to shrink the size and weight of Nano-Hyperspec is valuable by itself, it does allow the user more room and available payload to carry other instrumentation. Hyperspectral combined with LiDAR and thermal imaging is an extremely valuable package that is made possible thanks to the overall size/weight reduction of Nano-Hyperspec and the embedding of the data storage/management capabilities (which were contained within a separate enclosure previously).
Hyperspec III software gives users full control over data acquisition, sensor operation, and datacube creation in ENVI-compatible format. Hyperspec III also works in full conjunction with the GPS that can be paired with the sensor as an available Airborne Package. In this optional package, customers are able to take advantage of real-time computation of inertial enhanced position/velocity, ~161dBm tracking sensitivity, accurate 360-degree #D orientation output of attitude and heading, correlation of image data to GPS data, and much more. During post-processing, the Airborne Package also effortlessly handles radiometric calibration and conversion as well as orthorectification.
Growth Markets Require Solid Industry Background Across Commercial and Defense Markets
Fitchburg, MA – June 6, 2014 – With a rapid expansion of international business, Headwall Photonics announced today that Tom Breen has joined the Company as Director of Global Sales. Tom brings with him significant experience across many of the end-user markets served by Headwall. He will be responsible for managing Headwall’s growing worldwide sales activities and strategic opportunities for hyperspectral and Raman imagers as well as the Company’s OEM integrated spectral instrumentation.
Prior to joining Headwall, Tom held executive leadership positions at UTC Aerospace Systems where he was responsible for sales and business development of airborne and hand-held products. He also served as Vice President of Sales and Marketing for General Dynamic’s Axsys Technology Division in Nashua, New Hampshire. Other senior management positions at L-3 Communications, BAE Systems, and Lockheed Martin provided Tom with the background that will allow Headwall to grow its business in the hyperspectral imaging market.
“We are thrilled that Tom has joined our team,” said Headwall CEO David Bannon. “His background complements our commercial growth plans seamlessly and he will be a terrific asset in tackling a market that is experiencing very robust growth. Tom has had significant success in building high performance sales teams coupled with exceptional customer relationships.”
“I am very excited to be joining Headwall at a period of tremendous momentum for the Company and the industry,” said Tom. “As a leading supplier of spectral instrumentation, Headwall is uniquely poised to expand and deliver hyperspectral sensors and OEM instruments for remote sensing and in-line applications.”
Headwall’s award-winning Hyperspec and Raman imagers are used in commercial and military airborne applications, in advanced machine-vision systems, for document and artifact care, for plant genomics, in medicine and biotechnology, and for remote sensing. A unique differentiator for the Company is Headwall’s patented all-reflective, aberration-corrected optical technology that is fundamental to every system it produces.
Tom is a published author, with numerous works produced for IEEE, SPIE, and AAAE. Tom’s educational background includes MBA and BSEE degrees from Northeastern University in Boston.
Hyperspectral imaging sheds new light on prized Martian rock specimen
Scientists have forever been fascinated with space. What’s up there? Does life as we know it exist elsewhere? Is there any other celestial body like earth? While these questions might lack solid and precise answers, it’s not for lack of trying. Knowledge often comes not from massive ‘Ah-HA!’ moments, but from smaller discoveries. When stitched together, these jewels of learning present a useful mosaic for future scientists.
A two billion-year-old meteorite—officially named NWA 7034 but nicknamed Black Beauty by scientists—recently crashed into the Sahara desert. It was found by scientists in 2011 and determined to be of Martian origin two years later. The geologic history of Mars has always been a fertile source of exploration given the never-ending interest in this relatively nearby yet mysterious planet. While exploring the Martian landscape provides a wealth of scientific data, this meteorite has itself been a goldmine of information. Why? Because it sheds light not on the Mars of here-and-now, but on what we believe happened 2.1 billion years ago to its geologic interior and surface.
The Black Beauty meteorite was lofted off the martian surface by a large impact, an explosive geologic event. The intrinsic value of the rocks can be appreciated mostly because they carry a snapshot of what the conditions were like on Mars at the moment the impact occurred. The Mars of today is fascinating, yes, but to have a sample of Mars from 2.1 billion years ago is more fascinating still. Indeed, Black Beauty is significantly older than almost all other Martian meteorites yet found.
In early 2014, a Brown University research team led by Dr. Jack Mustard and graduate student Kevin Cannon temporarily acquired a slice of Black Beauty from Dr. Carl B. Agee, Director of the Institute of Meteoritics at New Mexico University. Brown University analysis included hyperspectral imaging using Headwall’s VNIR (380-1000nm) and SWIR (950-2500nm) sensors to extract a wealth of meaningful spectral data. "We were really presented with a one-of-a-kind specimen in Black Beauty," noted Dr. Mustard. "We wanted to learn as much as we could and add to the body of geologic knowledge already accumulated."
The team paired the two sensors in Headwall’s 'Starter Kit' configuration, which comprises a moving stage, necessary and proper illumination, and full software control to manage the collection and post-processing of the incoming data. "What we saw as we ‘unpacked’ the data is that Black Beauty is rich in information that give us a clue as to what Mars was like over two billion years ago," said Cannon. "While rovers on Mars today are extracting important new data, to have an actual sample that we can analyze with our most sophisticated instruments is exciting."
In the adjacent hyperspectral image of Black Beauty, features become clear. The mineral feldspar shows up as green, and the mineral pyroxene comes out as yellow/red. "These two minerals make up most of the Martian crust, so it's exciting that we can see them and map them out spatially in the data," said Cannon.
There are a few characteristics of hyperspectral imaging that make it perfect for this sort of work. First, it is a non-invasive technology. That is, no samples are harmed or even touched. This is crucial, and the non-invasive nature of hyperspectral imaging lends itself not only to the study of Martian rocks like Black Beauty, but also the field of fine arts, artifacts and antiquities. Museums and collection-care experts are themselves seeing the value of hyperspectral imaging because of the amount of new information that can be collected non-invasively.
As a scanning technology, hyperspectral imaging is designed to ‘see the unseen’ and unlock the answers to challenging questions. There are numerous ‘imaging’ and ‘scanning’ techniques available to the scientific research community, but none possess the vast spatial and spectral information collected by Headwall’s instruments. "What we have been able to do is successfully introduce a brand-new tool into our toolbox and prove its value," said Dr. Mustard. "We saw things in the VNIR and SWIR spectral ranges that no one has seen before, and our overall body of knowledge is more expansive because of it." Hyperspectral imaging collects ALL the spatial and spectral data within the field of view, not just some of it (as is the case with multi-spectral).
And what about closer to home, here on earth? Hyperspectral imaging is becoming more mainstream and affordable so that research entities like Dr. Mustard’s group at Brown can tackle projects like these more readily than ever. Graduate student Rebecca Greenberger has done similar hyperspectral analysis on rock and geological formations that many of us drive by without glancing twice. "There’s a rock formation behind a Target store in Connecticut that is just loaded with incredible geological samples," said Greenberger. Many of those collected rock specimens have themselves been scanned with Headwall’s hyperspectral instruments, yielding spectacular results and new information about the geological history of our planet.
Under cloudless skies in Ontario recently, Headwall achieved a very notable milestone: we became the first to fly both hyperspectral and LiDAR aboard a small, fully integrated handheld UAS. The test flights not only verified the reliable airworthiness of the system but also the ability to collect valuable hyperspectral and LiDAR data in real time.
Integration is key, because all of this specialized data-collecting instrumentation needs to fit the payload parameters with respect to size and weight. With UAS systems shrinking in size and weight, payloads need to follow suit. As prime contractor for this complete airborne system, Headwall is able to get end-users up and running quicker than ever. Time to deployment is reduced by months thanks to the work Headwall is doing to engineer optimized solutions that meet specific remote-sensing needs.
“The variety of applications for this type of integrated airborne system are numerous,” said Headwall CEO David Bannon. “Precision agriculture is a key one we’re seeing on a global scale, but geology, pipeline inspection, environmental research, pollution analysis are others.” Today’s UAS is smaller, lighter, and more affordable than ever, which makes it a perfect platform from which to carry precise imaging instruments such as hyperspectral and LiDAR. “We’ve always been a pioneer in the area of small hyperspectral sensors for just these kind of deployments,” noted Bannon. “Our strength comes from understanding what our users want to do and then engineering a complete airborne solution that meets that need.”
Chris Van Veen, marketing manager at Headwall, was on site to record and document the test flights. “A fully integrated package like this represents a new frontier for remote-sensing scientists who now have an airborne research platform that goes wherever they do,” says Chris. “Watching this fly and collect data in Canada was a thrill because it was visible testimony to all our integration work.”
The entire payload aboard this particular UAS is less than ten pounds, which includes hyperspectral, GPS/IMU, LiDAR, and computing hardware. Besides making sure these elements are small and light enough, the challenge of integrating everything with an eye toward battery lifetime is also Headwall’s to manage. “We know our remote-sensing users have very important work to do, and they need sufficient power not only to fly but also to operate the instruments,” said Bannon. One way to meet this challenge head-on is to make sure the hyperspectral sensor provides a very wide field of view with precise imagery from one edge to the other. “If you can assure outstanding image-collection across a wide field of view, and then provide orthorectification of that data, you’re covering more ground for each flight swath.”
Fundamental to accomplishing this is Headwall’s approach to optics, which is both simple and elegant. “Our diffractive optics approach uses no moving parts, which, in an airborne application, means robustness and reliability,” said Bannon. Inside each Micro-Hyperspec sensor is a precise and small holographic diffraction grating that manages incoming light with exceptional fidelity. These sensors are ‘tuned’ for the spectral range of interest to the user. “Depending on what the user wants to ‘see,’ he may need a VNIR sensor that operates from 380-1000 nanometers,” said Bannon. The spectral signature of a certain disease condition on a crop tree will determine the spectral range of the sensor, for example. Headwall has also introduced a wideband VNIR-SWIR sensor package that covers from 400-2500 nanometers. This co-registered hyperspectral instrument will be very popular with users who need broad coverage but need a small, light, and affordable instrument to do it with.
The following video will give you a peek into how flight testing went in Ontario.
One of the things we’re seeing at Headwall is the proliferation of airborne applications. Multispectral suffers a bit with respect to hyperspectral (a handful of bands versus hundreds), which is why hyperspectral is winning the day.
One reason is instrument affordability. Multi-million-dollar hyperspectral sensor programs might have flown (literally and figuratively) in the military world, but not in precision agriculture or with universities. Budgets are smaller, and that money has to be spread among not only the sensor but the UAV and everything in between. This is where small, entrepreneurial companies like Headwall shine, because everything in between can mean LiDAR, GPS/IMU technology, application software, data processing, and so much more. We understand hyperspectral imaging better than anyone, and our focus has always been to better that technology while driving costs lower. This is the essence of commercial-off-the-shelf (COTS), where highly specialized military instrumentation finds a home all across industry and academia. With respect to Headwall, COTS implementation means smaller, lighter and more affordable sensors that are easier to use yet just as optically precise as their multimillion-dollar military counterparts.
Second, you cannot go a day without seeing stories about UAVs. Fixed-wing designs like those from AGX and PrecisionHawk are crowding the skies along with multi-rotor helicopters like Infinite JIB and AIBOTIX. These are much more than hobbyist playthings and are perfect for scientific reasearch duties. They have excellent range and payload-carrying characteristics, and they are stable aloft. From mineral exploration and agriculture to petroleum and pollution control, UAVs are everywhere it seems. And everyone takes notice when household names like Facebook, Google and Amazon decide that the UAV is going to be instrumental to their future success. Much of this might sound fanciful and far-off, but it is happening now. Court challenges are being won, and while care needs to be taken on how regulations are drafted and enforced, no one doubts that the UAV is not only here to stay but will become commonplace.
Obviously, UAVs simply take up airspace unless they are doing good work. And largely, we seem to hear about bad things happening when mention of UAVs (and drones) is made. But stop and consider for a moment how a famine-stricken area can be made crop-fertile thanks to hyperspectral data that a UAV-mounted sensor can collect. A scientist will know about disease conditions with enough time to prevent damage by skimming the treetops and looking for anomalies that become ‘visible’ through hyperspectral imaging. A farmer will know where to plant and harvest…and where not to. Crop stress will be seen long before it becomes a worry, and the amount of wholesome and nourishing food planted in areas once thought impossible will blossom. In short, small and light UAVs are affordable for the people who need to use them. They can be flown in areas that vehicles and humans cannot yet reach, providing a window of research never available to scientists before.
As we see the proliferation of UAVs capable of carrying sensor payloads, it is important to understand how everything goes together. Here, Headwall is taking a leading role. Many mistakenly believe that slapping a sensor onto an octo-copter is all they need to do. But making sure everything works the way it should aboard a flying, unmanned vehicle is another challenge altogether. How much ground do you need to cover, and do you have enough battery power to do it? How much hyperspectral data do you need to collect, and do you have the computing and storage horsepower to make that happen? What are you looking for, and what spectral ranges are those things in? How do you ortho-rectify the data during post-processing? And how do you use the science of ground-truth as it relates to airborne hyperspectral imaging? This last consideration is hugely important, because the collaboration of airborne hyperspectral and ground-truth delivers the best possible accumulation of data. Headwall and ASD have even authored a 12-page whitepaper on the relationship between airborne hyperspectral data and ground-truth techniques.
When it comes to hyperspectral imaging, it isn’t always about the hardware. Before users even get to the stage of specifying a sensor instrument, they need to ask a few questions:
- What do I want to look at?
- How am I deploying the sensor?
- What is the spectral range of what I’m looking at?
- How far from the object will I be?
The answers to these questions will lead to an informed decision about the kind of sensor that’s best, the kind of lens it will need, and how small and light the sensor needs to be. At Headwall, we’re helping customers sort through these questions and considerations every day. We make on-line tools available that make instrument specification easy. With the answers to a few simple questions, the overall application-specific design of a hyperspectral instrument is well within reach. This means quicker time-to-deploy for customers who have challenging scientific questions that need answers.
One of Headwall’s newest tools is the Field-of-View (FOV) calculator. This tool collects a few important user-defined parameters to arrive at several what-if scenarios. The first parameter is distance from lens to object. In an airborne application, the distance would likely be measured in meters. For lab-based or in-line deployment, it might only be centimeters. The second parameter is the wavelength, which can be UV-VIS (380-825nm) all the way up to SWIR (950-2500 nm). Knowing the spectral signature of the item of interest will point you in the correct direction.
The calculator will take this information and combine it with choice of sensor and lens to arrive at useful data for the customer. In this case, we see that for the parameters and options chosen we are given the number of spatial and spectral channels (1004 and 335 respectively). We’re also given the linear and angular FOV, the instantaneous FOV, and the spectral resolution. In an airborne application, the linear FOV can be thought of as the flight swath. The wider the better, because the aircraft or UAV will be able to collect full hyperspectral information with fewer passes over the ground.
Spectral libraries are common starting points for defining where to look along the spectral range. The spectral signature for everything from plants and crops to minerals and petroleum is known or catalogued. While everything has its own signature, the real strength of hyperspectral imaging is to discriminate and classify. So while the sensor can actually ‘see’ everything, it is tuned to look for things that may resonate at 900 nm or 1900 nm for example. A disease condition on a fruit tree may be impossible to detect by any visible means, but it will resonate quite clearly when seen with a hyperspectral sensor.
Customers come to Headwall regularly with certain ‘needs.’ A crop scientist may want to analyze the soil from an airborne UAV. Another may want to adopt hyperspectral imaging along a high-speed food processing line to see and remove foreign matter. A third may be a museum preservationist interested in understanding the artwork and artifacts under their care. But in all cases, the first question is: What do you want to see?
As the market for hyperspectral sensing technology moves forward and advances, Headwall’s Application Engineering team has been able to gather a rare view into the past through the hyperspectral scanning of some of the most important historical artifacts and papers in the United States. For the first time ever, hyperspectral VNIR and SWIR imaging was conducted on key historical documents from the US Civil War period.
By working collaboratively with the researchers in the Cornell University Division of Rare and Manuscript Collections and the Cornell Johnson Museum of Art, Janette Wilson and Kwok Wong of Headwall’s Application Engineering team spent a few days conducting VNIR and SWIR hyperspectral scans of some of the most important artifacts held by Cornell University. Of particular interest was the hyperspectral scanning of the University’s collection of original Lincoln documents signed by president Abraham Lincoln during his presidency. This collection included the Gettysburg Address (seen at left), the Emancipation Proclamation, and the 13th Amendment to the Constitution.
The scanning of documents and artifacts with hyperspectral imagers is particularly well suited for the purposes of both 1) research and 2) for establishing a baseline of spectral/spatial information for monitoring change in the artifacts to better preserve objects of cultural heritage.
For a couple main reasons, hyperspectral imaging is particularly appealing to collection-care experts. First, and probably most important, is that the technology is non-destructive. The instruments don't interface with the documents and the lighting is called 'cold illumination.' That is, there is no risk of themal damage to the items under inspection. Second, previously unseen features immediately 'come to light' when viewed hyperspectrally. Note the image below, which represents a stamp on the Gettysburg Address that cannot be seen visibly but can when looked at within the VNIR and SWIR spectral range. Collection-care experts are fascinated by unseen features, which can be used to build the body of knowledge with respect to documents or artifacts.