New feature extraction tool for lidar

Exelis releases E3De amid rebranding effort

by Sam Pfeifle | November 28, 2011

BOULDER, Colo. – Just as its parent company was finalizing a split into three publicly traded firms, ITT Visual Information Systems, now Exelis Visual Information Solutions, released last month its first foray into lidar processing software that delivers automated feature extraction, E3De. Building on TLiD technology the company acquired, along with its own ENVI geospatial image analysis software and IDL programming language, this new product is aimed at letting users quickly pull out features like power lines and buildings from lidar data.

“Its specialty differentiator,” said Exelis solutions engineer Pete McIntosh in an interview with SPAR, “is that we focus on the extraction of features from point clouds … We’ll pull out buildings, power lines, trees, power poles, and instantly convert those into vector-based formats like shape files, so that they’re immediately useful in your CAD applications.”

Ideally, the information comes in LAS format, McIntosh said, as “there are inherent data structures in that format that are nice; it can give you RGB for a point, and it automatically has the lat-long built into it, but ASCI is fine, too. But that means the user has to do a little more massaging when you’re importing.” Other data formats can be made to work as well.

While the algorithms and math that go into making E3De work have to remain opaque, the potential applications are starting to become clear, most of them derived from airborne lidar data (though the company is exploring applications with terrestrial-generated point clouds).

“The utilities market is obviously an easy victory,” McIntosh said, noting the demand for lidar data aimed at showing compliance with North American Electric Reliability Corporation standards regarding vegetation around transmission lines. Also, “there are quite a lot of military applications that make use of lidar, especially for situational awareness. They do actively fly lidar instruments over theaters and our tools are being warmly embraced.”  

The development of solar maps, like this one from New York City, documenting the suitability of different locations for solar panel placement, is an intriguing market of E3De, McIntosh said. Exelis even performed a quick study for the city of Toronto, looking at which buildings would be suitable for the “green roof” movement, essentially spots where grass could be planted on top of buildings.

“We did a very quick and easy characterization of all the existing rooftops,” McIntosh said. “With very little effort we had a map of greenroof suitability, which could be of use to planners.” And, for that matter, a great sales map for those doing the green roof conversions.

Already collected data might even prove to have new utility, assuming the lidar has been collected at a high enough density. “You’re not going to get power lines if you’re only collecting at one point per meter,” he noted.

On some level, though, Exelis expects users to come back to them with applications the company hasn’t yet thought of. The software is good for creating quick fly throughs and visualizations and the number of features it can extract will increase in coming versions.

Noting that many companies might be sitting on lidar data they don’t even know what to do with, McIntosh postulated that as companies like Exelis create better ways to extract intelligence from point clouds, people will discover new ways to use point clouds. “We can render it and make it look really awesome,” he said, “but they don’t just want to look at it. They want information from it. That’s going to be the ultimate evolution of where lidar goes: The amount of information we can extract from a point cloud.”