Drones have become a quick and relatively inexpensive way to capture on-the-ground images of geographic areas, offering valuable information to soldiers and others. But those images have a drawback: They often cannot accurately be compared with images taken previously—images that generally were collected using different standards and formats—in order to determine how the terrain has changed over time.
Researchers with the National Center for Airborne Laser Mapping (NCALM) at the University of Houston (UH) are using a $1.89-million grant from the US Army Corps of Engineers to create a set of algorithms that would allow users to more-precisely align data sets collected at different times and reliably estimate changes between images captured at different times.
Craig Glennie, principle investigator with NCALM and associate professor of civil and environmental engineering at UH, said just-in-time data collected by drones can be helpful on the battlefield or in other circumstances. "More and more, people want to take older data and compare it to newer data and see if anything has changed," he said. "And there's not really a mechanism to do that with any confidence."
The main investigators on the project are Preston Hartzell, an assistant research professor at NCALM who is also involved in a similar project funded by the National Geospatial-Intelligence Agency, and Glennie.
The most complete data sets are collected by aircraft flying over an area, which allows researchers to use more powerful mapping equipment and to access GPS satellite data. The combination of more-powerful equipment and GPS data makes those data sets more precise, Glennie said.
Drone-collected data offers a different advantage—drones are small, nimble, and less intrusive. But they aren't able to access GPS data as accurately and, because the equipment they carry is smaller and lighter, they can't collect as much information.
To allow researchers to reconcile disparate data sets, Hartzell and Glennie will develop new algorithms—Glennie describes them as a new tool set—that can accurately highlight data signaling changes while ignoring extraneous or inconsequential data.
Glennie said the results will be useful in mapping natural disasters, as well as for military applications, including to help Houston and other cities accurately gauge the extent of flood damage following a hurricane or other major storm.
The resulting tools and methodology will be shared with other researchers through the open-source Point Data Abstraction Library.