If it feels like drones are everywhere these days, it’s because they are. Nearly 3 million drones regularly fly over American skies alone, according to the Federal Aviation Administration. With corporate behemoths like Amazon planning to shift in the future to a system that includes drone delivery, some industry experts are forecasting there will soon be more drones in the sky in one day than planes in a year. Globally, the market for drones is expected to double, from $30 billion this year to more than $60 billion by 2025.
But beyond just the private sector, drones have increasingly compelling uses for international development as well. Major players like the World Bank, USAID, Unicef, the UN, and the World Economic Forum are exploring the “gamechanger” aspects of drones, from mapping disaster areas and delivering life-saving medicine, like COVID-19 and HIV vaccines, to obtaining localized climate data.
Here at AidData, geospatial scientist Dr. Kunwar Singh is spearheading multiple projects that incorporate drone technology or satellite imagery. One current project on his team’s roster is a geospatial impact evaluation (GIE) funded by the Millenium Challenge Corporation that uses satellite imagery to identify how irrigation and better access to water has impacted productivity in Lesotho. The team is also seeking to incorporate drone-captured data for use in other future projects.
Dr. Singh, who is also an affiliate faculty member at the Center for Geospatial Analysis at William & Mary, joined AidData’s Research and Evaluation Unit (REU) in 2020. He has more than 15 years of experience working with remote sensing data and previously worked for the United Nations Development Programme; he holds a PhD in Forestry and Environmental Resources from North Carolina State University. In an interview, Dr. Singh explained to us how the REU is developing methods for translating raw data from drones into tangible measures of economic growth.
“Crop diversity was a key measure for us to track in this project. When more water is available, farmers are better able to diversify their crops,” said Dr. Singh. “We’re also exploring the potential use of drone data to look at how water resources and farm sizes change over time. More water means bigger farms, and this is the kind of outcome you’d hope for when evaluating a project that improves irrigation infrastructure.”
To measure water levels from so many different perspectives, Dr. Singh and the REU needed high-quality data. With high-resolution satellite imagery data from Landsat, Sentinel, DigitalGlobe, and Google Earth imagery, the researchers could turn nuanced geospatial information into a tangible metric of the effects of water access in society.
“The beauty of innovations like remote sensing data and drone imagery is that they let researchers create their own data when they find a gap in the field. That’s a powerful asset,” says Dr. Singh.
Dr. Singh is an expert in collecting and analyzing drone data. Along with Dr. Amy Frazier, director of the Spatial Conservation and Landscape Ecology (SCALE) Lab at Arizona State University, he co-edited and published a textbook on the topic, Fundamentals of Capturing and Processing Drone Imagery and Data, earlier this year. The book is now available in hardcover and ebook form.
“Sometimes you jump in first and start swimming later,” Dr. Singh said of the experience. “Amy and I started the book thinking it might be easy, but we soon discovered that there were virtually no courses available on the topic beyond the workshop level in many universities.”
This made finding contributors with relevant experience a grueling process. “Starting out, it was hard to know what area to focus on or where to find the experts. We searched Google Scholar and identified some people who were working on the topic, found out where they were, and contacted more than fifty people,” he said.
“We had such an overwhelmingly positive response that we were able to draw on experts from many different fields. This was perfect, as we thought presenting applications for drone data would be more appealing to students. This book brings everything together.”
“What we’re happiest about,” Dr. Singh added, “is that students can use the textbook to start by learning some of the basics, but then progress all the way to very advanced-level applications.”
The chapter design intentionally includes datasets that are freely available, so learners can work through presented problems as they would with an in-person teacher and examine the outcomes themselves. This public good was a particular criterion that Frazier and Singh wanted to ensure they were able to include.
As more universities start creating introductory classes on drones, resources like this textbook can help fill the knowledge gap and lead students all the way from their introductory courses up to advanced conceptual applications of drone imagery data.
While creating the textbook was a large driver in solidifying Dr. Singh’s specialty knowledge of drones, his first introduction to the topic came much earlier.
“I supervised a graduate student who was looking at using drone imagery data to help paper mills estimate the size of their woodchip inventory. The methodology ended up working so well that the drone estimates proved more accurate than their in-person surveyors. That was where I started with drones, and I am now working on my fourth project on the topic.”
Now, drone data and other high-resolution imagery is being used more and more across a variety of different fields. Questions that previously could only be answered with satellite and airborne data are now being revisited with much higher-resolution data at hand.
“We’re currently exploring applications where drone data can be leveraged to fill the data gap and bring projects to scale,” said Dr. Singh about future projects. “The prospect of using small-extent or local data from drones to answer questions at regional or even national scales is really exciting. Drone data can provide incredibly enriching information.”
Alongside the geospatial research teams at AidData, Dr. Singh is preparing to field a handful of concept notes, applying his knowledge and expertise with high-resolution data primarily to the fields of agriculture and the environment.
“One thing we’re looking at is livestock monitoring. In some countries, livestock is people’s primary livelihood, but having enough livestock requires knowing how much grass you have to fulfill the forage need.” With remote imagery, researchers can monitor how much livestock there is and the grazing capacity of the field.
“We’re also looking at water—in many countries, especially those economies reliant on agriculture, water is a perennial problem. There isn’t enough of it throughout the year, or there is flooding in the rainy season and droughts during the dry season,” Dr. Singh explained. “To more effectively grow crops under these conditions, in Niger and places like it, traditional farmers turn to pit farming, also known as ‘zai’ or ‘demilunes’, after the half-moon shape of the pits.”
But there is a knowledge gap when it comes to where pit farming is used and to what effect. Detection of pit farming itself is lacking. “We need to figure out where these pits are located in the first place, and if they are as impactful as we believe them to be, before we can begin to recommend implementing them in new locations,” said Dr. Singh.
Dr. Singh and the REU are also engaged in a crop monitoring study, using high-resolution imagery to detect extreme weather conditions, pests, and pathogens and then monitor the effects of these phenomena on crop growth. One applied advantage of this study is that crop insurance can only be claimed if the farmer can prove what happened to the crop. Having this information available would help insurers more accurately assist impacted farmers.
A final project in the works would look at the impact of water on crops. “In this case, we’d be looking at rice, which is an enormously important staple crop across Asia and parts of Africa. Droughts and floods are common problems when growing rice, so we want to help identify flood-resistant paddy crops to hopefully improve crop yields,” Singh stated.
Having access to drone data for monitoring seasonal changes, especially with respect to questions focused on agriculture and the environment, could be invaluable to continued efforts across these many avenues.