Multispectral LiDAR: Deserts, Rivers, Forests, More
During the 2016 MAPPS 2016 Winter Conference, LiDAR technology went home the winner. The overall grand award went to the Teledyne Optech Titan, a multispectral LiDAR sensor, which also won the Technology Innovation category. The innovative three-wavelength system had spent about a year being utilized in real-world operations before being recognized. So far its track record indicates that multispectral LiDAR will be of great value to scientists and commercial users for a variety of applications going forward.
Demand for Multispectral Capabilities
“Originally LiDAR was seen as a means of obtaining elevation data, but there is a growing interest among researchers in accurately mapping water/land interfaces and identifying species by spectral characteristics, which requires multiple laser wavelengths,” says Juan Carlos Fernandez Diaz, senior researcher at The National Center for Airborne Laser Mapping (NCALM).
NCALM, at the University of Houston, is a science research center funded by the National Science Foundation. Its primary function is providing research-quality LiDAR data to support the scientific community. To meet the needs of its users, NCALM requested a new LiDAR sensor that could collect high-resolution topographic data and quality bathymetric data in one pass, as well as provide enhanced spectral capabilities for more automated land cover and target classification. Teledyne Optech agreed to develop a sensor that could meet these criteria, and the result is the Titan multispectral LiDAR sensor. It got its trial in the harsh environment of the McMurdo Dry Valleys in Antarctica.
“It has taken a while for people to realize that intensity values collected with LiDAR are very useful, particularly in classifying land covers and identifying transition areas along rivers and coastlines. We were not looking for another deep-water bathymetric sensor — we needed one sensor that could provide high-resolution bathymetric and topographic information and return intensity at multiple wavelengths at the same time.”
“The Titan is like a Swiss army knife — it is multi-purpose so it doesn’t have the highest performance for one specific application or the other, for example compared to a bathymetry-only sensor, or compared to a high-altitude topographic sensor. Its capabilities are unique,” Fernandez says. “Depending on how you configure the flight plan, you can accomplish almost 100-percent illumination of the terrain in a single pass, which can sometimes be a metric more revealing than a simple point density.”
Multispectral analysis is relevant for commercial companies also. Foresters continue to search for a cost-effective method to classify species automatically using both spatial and spectral characteristics. Other users interested in vegetation identification include utilities responsible for keeping rights-of-way along pipelines or power lines clear.
Due to the productivity and cost advantages, there is hope that in the future land cover maps could be produced using data only from LiDAR sensors with enhanced multispectral capabilities. “With LiDAR, users have longer flying opportunities with no concern for sun angle and lighting, and LiDAR data is not impacted by solar shadows,” Fernandez says. “We are very interested in the further development of multi-wavelength high-density LiDAR sensors to support continued research in this area.”
All-in-One System Design
The Titan sensor includes two infrared channels at 1064 and 1550 nanometers and a third channel at 532 nanometers, commonly used in bathymetric sensors. This combination of wavelengths is able to capture land and water, as well as the transition areas in between with an effective ground sampling rate of 3-by-300 kHz for a total of 900 kHz. This flexibility allows users to collect mixed land cover areas in a single pass, rather than requiring separate flights with a topographic sensor and a bathymetric sensor. The Titan system also includes a metric digital camera to provide high-resolution RGB/CIR image overlays if necessary.
“To address NCALM’s interest in combined topo/bathy capabilities, Teledyne Optech created a multispectral LiDAR sensor that leverages principle performance expectations from a variety of application requirements to enable a broader application approach and system design,” says Michael Sitar, business manager, Airborne Mapping Solutions at Teledyne Optech. “For example, high laser sampling rates with narrow beam divergences were combined with a programmable scan field of view (FOV) to enable superior modeling of complex underwater terrain morphology. The result is a multi-use sensor, which can be purposed for more than a single application space, offered at a price similar to that of a single-use sensor design.”
Titan’s three wavelengths, collected simultaneously, produce data suitable for topographic mapping, land cover classification, shallow water bathymetry, environmental modeling, and forest inventory and vegetative classification. “Since we’re collecting each wavelength independently, the data can be operated within a GIS environment as layers of information, or combined and analyzed as discrete points,” Sitar says. “Some users fuse the independent intensity values (i.e., target amplitudes sensitive to the emitted wavelength) into a base layer and apply an RGB color palette to each wavelength.”
Additional features make the Titan flexible for multiple applications. The RGB/CIR camera is a complementing technology, adding additional metadata and the opportunity to create ortho-imagery, which can be used to validate land cover and species classification accuracies. The sensor also includes full-waveform recording capability for each wavelength, for those seeking to determine target detection confidence levels below the water’s surface. In forestry, waveform data offers additional information from the canopy rather than just purely XYZ measurement. Discrete data enables faster throughput, but Optech sensors offer the option to capture both discrete and waveform data when desired.
To provide maximum application diversity, the Titan’s standard configuration includes three independent channels that enable three times the structural information and detail than that of a collinear design, whereby all three beams occupy the same laser footprint on the ground. A collinear design results in far lesser point density; however, this type of customization is an option if there is interest in an application-specific beam configuration.
Diverse Applications Benefit from Flexibility
NCALM’s first major project with its new Titan unit involved mapping more than 3,500 kilometers² of the McMurdo Dry Valleys in Antarctica to provide data for topographic change studies. The dry valleys are the closest thing to Mars on Earth, with overlapping coverage of snow, dirt and ice. “It’s basically a frozen desert,” Fernandez says. “A single wavelength (1,550 nanometers) laser would have a disadvantage because snow and ice are not very reflective at that wavelength, while the Titan’s 1,064 nanometers will pick it up. Contrasting land covers have low reflectance in some wavelengths and higher in others, so having three wavelengths makes the Titan more flexible and more valuable in these situations.”
“This year we also successfully mapped the ancient city of Teotihuacan, in Mexico’s high desert,” Fernandez says. “By overlaying the 3D elevation model with the Titan’s three wavelengths, we can see things not visible with only a single wavelength. The resulting images of pyramids and the surrounding areas are like works of art.”
The Titan’s three-wavelength capabilities are also being utilized by Norwegian company Terratec. An expert in land surveying, laser scanning, aerial photography and cartography, Terratec collected multispectral LiDAR data over two 35-kilometer-long sections of the Gudbrandsdalslågen River and its tributaries in September 2015. The rivers typically flood twice each year as a result of melting snow and glaciers, as well as heavy rains.
To develop a hydraulic model to study the water volume and flow of the rivers, the Norwegian Water Resources and Energy Directorate (NVE) required both high-quality topographic and bathymetric maps. NVE is using the model to identify and test options for flood mitigation, which include building new protections and/or expanding existing ones, building flood tunnels, opening up old flood canals, and dredging and excavating masses from the riverbed.
“The Titan sensor is attractive to Terratec because it can be mobilized efficiently for both topographic projects and shallow water bathymetry projects,” says Leif Erik Blankenberg, technical manager, Terratec. “We also saw a potential for this multi-beam sensor in forestry applications. There are distinct advantages to having three wavelengths for classification purposes.”
“Airborne laser scanning has recently revolutionized forest inventory in many countries,” Blankenberg says. “For example, in Finland approximately 70 percent of forest area has already been inventoried using LiDAR; however, traditional LiDAR does not provide enough information to identify tree species. Digital aerial photography has therefore been used to separate the species. The introduction of Titan has awakened hopes for one-sensor, species-wise inventory. The first experiences by Terratec and others indicate that the multispectral information from Titan's three channels provides at least as much information as traditional LiDAR and digital aerial photography used together.”
The Future of LiDAR Technology
As scientists and commercial users continue to test the limits of LiDAR technology, its full potential will be realized. “LiDAR has become a very dynamic sector within the geospatial commercial industry, with many more sensor options available today than just a few short years ago,” Sitar says. “Clients are asking to do more with less. It was a logical progression to add more capability into a single sensor design to open up new avenues of collection opportunity and application development. I believe we will see continued development and evolution of LiDAR applications as the capabilities of sensors expand.”