New airport method takes flight
(Left to right): Photos, discrete lidar and waveform lidar point-clouds of common airport obstructions—a tree (top) and tower (bottom)—showing the achievable point density increase using full-waveform data. (large image) |
Does a tower near an airport need a flashing beacon on top? What sizes of planes can land on a given airstrip? Where can a busy airport add a new runway? A new method developed by UW-Madison engineers could quickly and efficiently answer these questions and others.
Airports need regular area surveys that map possible obstructions to help plan construction, tree maintenance and runway approach patterns. Typically, a surveying crew physically takes measurements on the ground, an expensive and time-consuming process. For his graduate studies in Civil and Environmental Engineering, Christopher Parrish chose to investigate another promising method of survey: airborne light detection and ranging or lidar.
Lidar works similar to radar, but uses laser light as its signal. As a surveying plane flies over the area that officials want to map, it sends out a laser pulse. Sensors on the plane detect the signal as the laser reflects off the surfaces it encounters. Then, engineers collect data from all the beams that scanned a particular point in space and map all the detected reflections in a scatterplot. Officials can use the resulting point clouds to determine the shape, size and location of obstructions.
The National Oceanic and Atmospheric Administration’s National Geodetic Survey (NGS) has researched lidar for the past decade. Traditionally, NGS has used discrete data, focusing on only the initial return of each laser pulse, or the “front edge” of the return signal. Now, systems can digitally acquire and save the entire laser return, a process known as waveform lidar—Parrish’s main interest. Waveform methods return much more information, creating scatterplots with an average of 252 percent more data points.
However, in processing the full return signal, traditional methods require a trade-off between resolution and noise. As filtering methods try to sharpen the signal to pick out individual points of reflection, they also amplify atmospheric noise, making it difficult to distinguish genuine signals.
While wrestling with these issues, Parrish sought advice from McFarland-Bascom Professor of Electrical and Computer Engineering Rob Nowak. An expert in signal processing, Nowak suggested Parrish try a different approach. “Traditional methods focus on what happened to the signal—the process that distorts it,” says Nowak, “but they ignore the physical characteristics of the signal itself.”
Nowak and Parrish developed a new workflow for processing data from waveform lidar, taking into account models for both distortion and signal characteristics. The approach resulted in a robust, reliable obstruction mapping method that addresses previous challenges while simplifying workflow. Parrish applied the new methods to lidar signals collected around the Madison area with great success. The two published their results in the May 2009 issue of the Journal of Surveying Engineering.
“This is a huge advance,” says Nowak. “It totally revolutionizes how well they’re able to do these automatic airborne surveys of airports.” Parrish estimates a 46-percent decrease in total obstruction survey completion time and a 38-percent decrease in human labor time, from the most recent NGS lidar obstruction survey.
“One of the things that makes this interesting is that it’s an application, but the basic principles are part of a very important area of signal processing from highly distorted, incomplete or noisy measurements. Lots of signals fit this kind of framework—MRI, for example,” says Nowak. “This is a great example of how those ideas can make a big difference in practical application.”