Beyond 3DMultispectral Optech Titan Opens New Applications for Lidar

A 5.817Mb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE

Lidar, as a tool for the remote sensing of the earth’s surface, has evolved into a relatively mature technology for conventional applications such as topographic mapping and 3D modeling. Selective wavelengths are often used to address specific application requirements, but these have largely been restricted to bathymetric surveying and water depth measurement. As a result, airborne lidar manufacturers have focused largely on performance and SWAP (size, weight and power) improvements, making their systems smaller, faster, more accurate, and more efficient. Such incremental improvements are both necessary and welcome, but to date, lidar systems have remained firmly in their niche as predominantly topographic or bathymetric surveying instruments.

The recent release of an Optech Titan multispectral lidar system breaks away from the traditional lidar paradigm of mere range measurement by exploiting differences in the target’s response to varying emitted laser wavelengths. By surveying simultaneously with three different wavelengths, Titan identifies not only the target’s position, but also the target’s unique spectral sensitivity to the emitted wavelength expressed as a target response amplitude (intensity), or reflectance value if radiometrically corrected. This new information content allows surveyors, researchers, and resource management professionals to derive information beyond what is possible for current applications, and opens up new possibilities and applications for lidar henceforth. Using three high-resolution active imaging channels with bands in the infrared and visible spectrum, Titan breaks the dichotomy of topographic versus bathymetric lidar systems, enabling a single-sensor design that can deliver superior results without the need for multiple sensor systems or separate projects.

Many of Optech’s technologies and innovations come directly from the needs of their clients. Optech’s ultra-compact Orion lidar sensor, for example, is the commercial result of the US Army’s requirement for a single, self-contained complete lidar system suitable for dual use on manned and/or unmanned aerial vehicles. The Optech Titan multispectral lidar system also originated from a customer’s unique requirements for a combined topographic and bathymetric sensor design that did not sacrifice resolution, depth performance, data precision or accuracy. With no such system available on the market, the National Center for Airborne Laser Mapping (NCALM) at the University of Houston contracted Optech to develop a new sensor, culminating in the commercial release of the Optech Titan.

NCALM has a specific interest in topographic and bathymetric mapping, which Optech previously addressed with separate sensor designs. These application-specific designs were optimized for one or the other, but could not excel at both simultaneously. Therefore, it was appropriate to consider a newer sensor design that combined the two capabilities into a single innovative solution that could handle topographic and bathymetric applications at the same time without compromise. Optech further extended this base requirement with the addition of a 3rd channel, enabling new applications beyond simply topographic and depth measurement.

Core to the Optech Titan design are two infrared channels at 1064 and 1550 nm and a third water-penetrating "green" channel at 532 nm. The beams are transmitted to the ground via a programmable, oscillating scanner with a userselectable FOV, which enables the user to significantly increase point density by narrowing the FOV, maximizing target resolution and detectability. The separate IR and green channels are imperative for accurate water surface measurement and land/water discrimination, and therefore accurate water depth measurement. Single-wavelength sensor designs have proven to be less effective in this regard, as they have difficulty accurately locating or measuring the shoreline interface and water surface.

The beams are currently configured with a 3.5 separation from nadir (i.e. 0, 3.5 and 7). The green beam is 7 off-nadir, preventing specular reflection off of the water surface that could hinder water depth penetration at altitude. This configuration is termed a "non-collinear" design since the three independent beams/wavelengths do not hit the same target simultaneously. Commercial user feedback thus far has indicated a preference for such a universal sensor that can accommodate a variety of application requirements. Furthermore, the effective ground sampling rate in this configuration is 3 300 kHz = 900 kHz. Users interested in classification opportunities can handle the three laser channels as separate 3D data layers with unique reflectance values sensitive to the emitted wavelength of the specific channel. Data is then interpolated to a continuous surface and standard image processing techniques are applied.

Optech is dedicated to providing sensors that meet clients’ particular requirements. For users interested in a specific beam configuration that is application specific (e.g. collinear designs where the emitted beams strike the same target simultaneously), Titan can be configured accordingly at the factory. In this configuration, the effective ground sampling rate is 1 300 kHz = 300 kHz.

By default, Titan uses a native discrete measurement technology that incorporates an on-board analog signal analyzer to capture high-precision range measurements and produce lower data storage volumes without affecting data resolution. This is the industry-standard approach for surveyors who desire maximum data precision and accuracy without the processing burden of voluminous waveform deconvolution during post-processing. As sampling rates get larger, this becomes a critical consideration in the context of timely deliverables. However, users that need to extract additional information can independently record the full waveforms of all three imaging channels and process them using Optech’s standard lidar processing software.

Effective and efficient sensor design is imperative to meeting system user needs. While Optech’s initial focus was to develop a single sensor design that leveraged multiple wavelengths, additional capabilities were incorporated for maximum collection efficiency and configuration flexibility. These include full gyro-stabilization for predictable point distribution and fully-embedded passive imaging options with the user’s choice of highresolution RGB, CIR, NIR, thermal and multispectral options.

Titan’s primary attractiveness is not only its application diversity, given its uncompromising approach to traditional topographic and bathymetric applications, but more specifically the opportunity to more efficiently classify target responses into unique landcover classes using active sensing techniques alone. When coupled with the inherent 3D properties of the lidar point cloud, multispectral lidar becomes a powerful new remote sensing tool. Whereas traditional single-wavelength lidar systems are very good at revealing where the targets are, they are less capable of determining what the targets are. Surveyors can certainly identify targets using single wavelength intensity (i.e. uncalibrated reflectance) or normalized reflectance values, but this only reveals the target’s spectral response along one wavelength. Passive sensors are another possibility, but they can be affected by terrain shadows and varying solar illumination levels, and cannot function at all at night. Multispectral lidar breaks the paradigm of coordinate measurement alone by allowing surveyors to compare a target’s unique reflectance response for each wavelength emitted, providing more detailed analysis and enhanced target discrimination.

For higher classification success, it is imperative that target responses (i.e. raw intensity values) are normalized by removing the radiometric biases associated with range, angle of incidence, atmospheric conditions, etc. The percentage of target interception within the emitted footprint also requires consideration.

Preliminary tests using standard image processing algorithms, such as an unsupervised maximum likelihood classifier used by Ryerson University, have produced exciting results. Test data collected by Titan over a suburban area of Toronto, Canada was used to try and identify six basic land cover classes (buildings, trees, roads, grass, soil and water). Two hundred random checkpoints were independently collected and verified to confirm classification accuracy. Based solely on the combined radiometrically corrected intensities of the three channels, Titan successfully classified targets with an accuracy of up to 69% using lidar reflectance alone. Adding a DSM generated from the highresolution point data as an additional band increased overall classification accuracy even further to 78%.

Higher classification accuracy may be possible by using alternative lidar wavelength configurations or by adding co-collected passive imaging data, but the potential is clear. While band combination and/or laser wavelength selection can obviously have an impact on what can or can’t be discriminated, Titan is certainly a very capable sensor for classificationbased applications normally associated with passive imaging sensors alone.

Time will tell what Titan is truly capable of doing, though NCALM is certainly putting their newly acquired sensor through its paces. Titan is currently deployed in the McMurdo Dry Valleys of Antarctica, collecting high-resolution multispectral lidar data aboard a Twin Otter aircraft operated by Kenn Borek Air.

At the time of this writing the NCALM team was communicating from McMurdo Station, Antarctica, where they are flying two 3.5-4.5 hour missions per day when the weather permits. The weather is not predictable like it is at home, and there have been a significant number of cancelled days, and many days where they are only able to fly one mission. Research Associate Darren L. Hauser shared the following:

"The goal is to map the McMurdo Dry Valleys (which are a pretty unique place in Antarctica) to study topographic change. There are many valleys, but two of the main valleys are Taylor Valley and Wright Valley… One thing I’ll say, the research community down here is VERY excited about what we’re doing. Everyone who’s seen the preliminary Optech Titan results and images has been enthusiastic."

Note: Optech thanks NCALM for the Optech Titan imagery from Antarctica they provided for this article.

Michael Sitar is the Business Manager of Optech’s Airborne Survey Division. With more than fifteen years of lidar experience within the airborne survey technical management team, Michael has past experience as a Field Operations Coordinator, Manager of Airborne Operations, and Airborne Products Manager for Optech.

A 5.817Mb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE