Shedding Light with LiDAR
Written by Peter Buxbaum
GIF 2012 Volume: 10 Issue: 1 (February)
More than anything else, one key circumstance has contributed to the growth in the collection, analysis and exploitation of light detection and ranging (LiDAR) data in recent years: U.S. forces have found themselves fighting in theaters in which they owned the skies, allowing the aerial overflights that collect LiDAR data in Afghanistan and Iraq to proceed undeterred.
From there, it was just a matter of time before military and industry imaginations took over, thinking up and developing new and better ways to collect, extract, analyze, exploit and apply LiDAR data.
LiDAR, a technology that has been around since the mid- 1990s, uses 1.064 nanometer wavelength laser light pulses to gauge distances by measuring the time delay between transmission of the pulse and detection of the reflected signal. A range finder mounted in an aircraft swings back and forth collecting data on up to 150,000 points per second, providing resolutions of one point per meter on the ground and one point per 15 centimeters vertically. The data returned by the LiDAR sensor provides location data on an x-y-z axis, referred to as a point cloud.
LiDAR has an advantage over some other geospatial technologies in that it provides accurate elevation data. LiDAR outstrips the capabilities of other sensors with its ability to pinpoint the location and elevation of surface elements such as buildings, trees and roads. Under the right circumstances, it can also detect hidden objects, for example by penetrating forest or jungle canopies.
But LiDAR’s true value as a military and intelligence tool, say the experts, comes when it is used in conjunction with data from other sources such as electro-optical, infrared and hyperspectral sensors to enhance the picture used by analysts, planners and commanders. The same kind of fused data is also being used in simulations incorporated into training systems.
“Changes have come to LiDAR in recent years,” said Matt Morris, director of product development at Overwatch Systems. “There are more providers out there building sensors. The costs to collect the data have come down dramatically. But the biggest development is that the Department of Defense and the intelligence community are using LiDAR in their daily workflow. When the defense and intelligence folks make something a priority, the effect is seen across the whole market space.”
U.S. forces in Afghanistan use the BuckEye system, which combines airborne LiDAR technology with digital color camera imagery to provide pictures to commanders and planners on the lay of the land. LiDAR elevation data supports improved battlefield visualization, line-of-sight analysis and urban warfare planning.
Fusing data from multiple sources increases the probability that features can be automatically extracted from the raw data and that an accurate situational picture will result. Automated feature extraction is a capability that allows software to recognize certain specific objects represented in LiDAR point clouds. Programming the software to be on the lookout for topographical features such as hills or man-made objects such as buildings, vehicles or power transmission lines allows those features to be separately and distinctly portrayed in the LiDAR image.
“The total is greater than the sum of its parts,” said Matt Bethel, manager of systems engineering at Merrick GeoSpatial Solutions. “The more data than can be put together increases the probability that a specific target is what you are looking for and decreases the uncertainty of that decision.”
“LiDAR data is becoming more critical and important for developing accurate and current visual databases used for training pilots, troops and tank commanders,” added Pratish Shah, director of marketing at Quantum3D. “Manned or unmanned air and ground vehicles take LiDAR scans of a given area, which represents the most current visual information of that region. Integrating that current information improves realism of virtual training systems. Current data can also be used to extend training to mission rehearsal, giving pilots and ground troops access to the most current visual data to train for specific missions.”
Storage and Dissemination
The large volumes of data generated by LiDAR sensors represent a potential disadvantage to the use of that data. Storing LiDAR point clouds is more complex than storing and rendering image files because the addition of a third dimension renders the task more computationally intense.
“Point clouds are actually nothing but a pile of x-y-z data,” said Oodi Menaker, marketing product manager at Israel-based Tiltan Systems Engineering. “The main challenge is to extract point cloud data, which you can then work with to describe the ground, buildings, power lines, trees, power poles and many other geographical features.”
Dissemination of these huge data files also presents a problem. “Once you collect massive amounts of data, how do you disseminate it on existing networks to a lot of concurrent users?” said Rudi Ernst, president of Pixia. “If you can’t access the data, what is the point of collecting it?”
The development of multi-sensor data fusion suggests the collection of raw data with multi-sensor packages. “Merrick specializes in multi-sensor collection and analysis,” said Bethel.
Merrick provides LiDAR capabilities to the National Geospatial- Intelligence Agency, U.S. Naval Research Laboratory, and U.S. Army Research, Development and Engineering Command, operating the aircraft and sensors that collect the data and develop products based on customer needs.
Sensor and analysis requirements differ based on mission and geography, according to Bethel. “A customer in Latin America was interested in identifying semi-submersible submarines that are used to move cocaine,” he said. “LiDAR is very useful in that effort, but you also need thermal and hyperspectral sensors to see whether you have a positive identification or a false hit.”
In Afghanistan, LiDAR is often used to help identify sites where IEDs may have been buried. “When a LiDAR sensor has passed over the same area multiple times, the data can be used for detecting changes in terrain,” Bethel explained. “An area where the ground has been disturbed may indicate the location of an IED. But it also pays to include a hyperspectral sensor which can detect the chemical signature of the explosive.”
Merrick utilizes software to visualize LiDAR point cloud data to develop products for customer consumption. The company’s Merrick Advanced Remote Sensing product is a Windows application used to visualize, manage, process and analyze LiDAR point cloud data. “Once the LiDAR data is processed it takes many steps to get to what the client is looking for,” said Bethel.
Multi-INT Focus
Northrop Grumman focuses less on geospatial intelligence in isolation and more on multi-INT in response to customer demands, said Sean Love, a company business development manager. “The real focus going forward is on taking geospatial data and pushing it to other intelligence types and ingesting other intelligence into geospatial,” he said.
The company facilitates the fusion of different data types through translators supported by a service-oriented architecture. Northrop Grumman has expanded its existing geospatial portfolio to include geospatial data acquisition, collection and processing of LiDAR, full motion video and persistent surveillance data, photogrammetric services, geographic information systems and analysis.
“All this is transparent to the user,” said Love. “The user doesn’t care what the application looks like. He just wants to be getting the right kinds of information. That is why we focus on getting actionable intelligence to analysts and warfighters.”
The company’s offerings support functions such as intelligence gathering and mission planning, routing and logistics, execution monitoring, physical asset tracking, exploration of what-if scenarios, data exploitation and analysis, highly integrated databases and sensor networks, and secure command and control systems.
“In 2011 and in 2012, we are in the process of automating these processes and making them a lot faster,” said Love. “We have condensed the process of transforming LiDAR point clouds to a topographical map from days or weeks to minutes. We are now focused on the first part of the process—organizing point clouds from raw LiDAR data.”
The challenge in the processing of raw LiDAR data is in the math, said Love. “It’s all about the algorithms and getting smarter about it,” he added. “It is doing the error correction and consolidation, especially if you are doing multiple collects. It takes a long time to crunch that information. It involves taking raw data and shaping it into something the user can do something with.”
Northrop Grumman’s efforts at automation are aimed at taking users largely out of the preprocessing routine and data correction process. “We can now load all of the data on a server and let the software crunch it,” said Love. “The user doesn’t have to sit there swapping out disks. Earlier processes had users on the lookout for data anomalies. Now the computer spots these mistakes and corrects them automatically.”
3-D Visualization
Quick Terrain Modeler, a 3-D point cloud and terrain visualization software package from Applied Imagery, was designed for use with LiDAR, but is flexible enough to accommodate other 3-D data sources, said Chris Parker, the company president. “Quick Terrain Modeler works with large 3-D data sets. It doesn’t matter whether it is LiDAR, synthetic aperture radar, or sonar. LiDAR happens to be the major 3-D source right now, and a majority of our users use Quick Terrain Modeler to work with LiDAR.”
A recent update to Quick Terrain Modeler, released last summer, provides greater speed and improved workflows. The tool’s workflow is called FLAP, for find, load, analyze and produce.
“Speed gains are achieved through optimization of analysis processes and by pushing more functions out to the graphics card,” said Parker. “Our latest release has a completely redesigned and intuitive interface, including a tool to keep the workspace organized and a mini-map to provide context when zoomed in, tilted or rotated. These gains translate to faster exploitation, production, briefing preparation, planning cycles and decisionmaking, and a shorter learning curve.”
Quick Terrain Modeler analyzes LiDAR point clouds and represents them as pixels to provide a digital elevation model (DEM). Users have the choice of working with a DEM or directly with the point cloud data.
Many users rely on processed LiDAR data to develop PowerPoint presentations. Quick Terrain Modeler enables that automatically. “Everyone up and down the chain in DoD needs to create PowerPoints,” said Parker. “When the complex analysis is complete, users will want to export that and share the information with others.”
Tiltan has developed a software program called TLiD, which “enables users to make sense out of the point clouds,” Menaker explained.
“We have automated the process of transferring point clouds to geographical features,” he added. “The latest improvements in the ability to depict precise features from LiDAR point-cloud data involve advances in the software algorithms used to process this information.” Tiltan has sold the exclusive rights to TLiD to ITT Exelis.
TLiD provides automatic extraction of features such as houses, trees and power lines; automatic full scene 3-D reconstruction; and output in a variety of file formats. It is integrated with a 3-D viewer.
“TLiD has the ability to develop and run complex urban scenes that include thousands of buildings and vegetation along with hundreds of moving objects,” said Menaker. “This is essential for MOUT [military operations on urban terrain] scenarios, including close air support and UAV operations.”
Disaster relief is another application of the analysis and visualization provided by TLiD, according to Menaker. “By comparing the pictures pre- and post-disaster, users can assess the status of bridges and rural roads.”
Overwatch has developed a tool called LiDAR Analyst, which was built as an extension to GIS products such as Esri’s ArcGIS. “It was ahead of the curve when it was first released in 2006,” said Morris. “Now it is picking up steam as LiDAR data is becoming more readily available.
“ArcGIS is one of the standard tools in the geospatial marketplace,” Morris added. “LiDAR Analyst dovetails directly into the ArcGIS workflow. Because we use standard formats, the output can be consumed by other applications such as Google Earth.” Overwatch recently improved its tools for LiDAR visualization to include automated feature extraction. The company is also working to expand its cataloging products to include the ability to search for LiDAR files.
LiDAR Fusion is Quantum3D’s most recent software application. “LiDAR Fusion is a visualization tool to help the geospatial intelligence community visualize point cloud data and automatically identify key objects such as buildings, vehicles and people,” said Shah. “For visual simulation applications, LiDAR Fusion can be used to extract information from a LiDAR scan and place that information into the virtual database used for training pilots, ground troops and tank commanders.”
Quantum3D uses LiDAR data in simulation and training applications to develop current and realistic visual databases for training military pilots. The visual data automatically extracted from point clouds can be imported into an OpenFlight database and can be used on any simulation platform, including Quantum3D’s Mantis Real- Time Scene Management software platform and Independence IDX real-time image generator platforms.
“Using current LiDAR data, for example, vegetation information is automatically extracted and placed into our flight simulation database,” said Shah. “Accurate placement of trees adds a level of realism as military pilots train. Pilots who regularly fly into military bases comment about tree placement in the virtual environment being very representative of the real environment.”
Data Access
LiDAR point clouds are dense with data, and this raises the question of how this is to be transferring across busy networks. “We are focusing on the data access piece,” said Pixia’s Ernst. “Dissemination is our core mantra.”
Pixia focuses on storing data in a way that allows for quick random access. “Pixia handles scalability on the server side,” said Ernst. “To have massive random access, the ultimate goal is to make a spinning disk perform like solid state memory. Our software boosts disk performance so that to accommodate hundreds and thousands of concurrent users.”
Pixia allows data to be accessed at the object, rather than the file, level. That way, users are able to access the snippets of data that they need rather than having to wade through an entire LiDAR point cloud file.
For Bethel, LiDAR’s future will see even greater data density. The frequency of the laser pulses on sensors will increase, allowing for both greater detail in small areas as well as capturing wider swathes of territory.
As multi-sensor data collection becomes more commonplace, Bethel expects that sensor packages will become miniaturized as well as modularized. “Right now it takes a lot of experience, knowledge, time and effort to combine sensors into one aircraft,” he said. “Future sensors will be more compact, more simplified and more rugged to better integrate them into unmanned aerial vehicles.”
“We see an increase in LiDAR use across both geoint and visual simulation community,” said Shah. “As LiDAR scanners are used repeatedly in unmanned air and ground vehicle environments, we see real-time LiDAR processing becoming more prominent. Getting LiDAR data from scan into a visual database faster and near real-time will continue to enhance the benefits of virtual training for mission rehearsal.”
For all of the developments associated with LiDAR, enhancement to collection and application of the data will only continue into the future. “We see LiDAR as a big growth opportunity for our business,” said Morris. “People are just scratching the surface with what they can get from LiDAR.” ♦
