Original post seen here, written by Valerie King at POB.
Photo Courtesy Quantum Spatial
As with many geospatial data acquisition approaches, the technologies that make LiDAR possible have experienced numerous advancements over the years — from hardware to software to application relevance to integration aptitude — with the aim to increase usability and decrease cost. With nearly 70 percent of respondents to the latest POB Laser Scanning Surveying Trends study indicating growth in demand for LiDAR services, POB decided to revisit the topic of LiDAR in general and examine its place in the geospatial world today. In recent interviews with representatives from Quantum Spatial, Riegl and Velodyne LiDAR, POB asked about the latest advancements in LiDAR, in addition to fundamental questions that newcomers to the remote sensing method may have.
Photo Courtesy Quanum Spatial
Terrestrial, Mobile and Aerial LiDAR
An important thing to know about laser scanning is that it is a very multifaceted data acquisition approach. There are numerous LiDAR techniques, which call upon a variety of hardware and software packages. For starters, there is terrestrial LiDAR, mobile LiDAR and aerial LiDAR.
Terrestrial data is captured from a stationary location, often using a tripod, with a scanner able to collect data within a 360-degree view of its location, explains Mark E. Meade , PE, PLS, CP, senior vice president of Quantum Spatial. He runs all data acquisition and international operations, and says LiDAR makes up about 65 percent of the company’s service offerings. He says terrestrial data is typically the most accurate point data and includes the highest densities. In his experience, terrestrial project costs, measured in cost per unit area, are typically higher than mobile and aerial LiDAR. Quantum Spatial typically uses terrestrial LiDAR for applications such as infrastructure inspection, as-built models and erosion measurements.
Mobile LiDAR, often referred to as mobile mapping, also involves the collection of point data from a ground-based platform, but unlike terrestrial, it utilizes a moving platform. Sensors are often mounted on an automobile and driven along pre-planned routes, Meade says, pointing out that mobile LiDAR can also be used on boats, ATVs and trains. “Mobile mapping is most often used in corridor applications, as opposed to wide-area collects. Many mobile projects are related to roadway, railway, airport, dams or riverine applications,” he explains.
Unlike terrestrial and mobile, aerial LiDAR is captured from an aerial platform that flies over the project area during the data acquisition phase. Meade says most projects today use a fixed-wing platform such as an airplane for acquisition, but that specialized applications of rotary-wing/helicopter acquisition continue to grow in popularity in cases when accuracy and point density are paramount. Airborne LiDAR is often the selected method for collecting high-resolution elevation data over broad areas, and the use of drones as a platform is an increasingly popular lower-cost alternative for projects requiring acquisition from closer up than planes and helicopters.
“The differences between terrestrial, mobile and airborne LiDAR generally fall into the broad categories of accuracy, point density, and project cost,” Meade says. “The added value for improved horizontal and vertical accuracy should be obvious and these technologies generally span the range from a few millimeters to 10 centimeters.”
The level of point density is highly correlated with a particular project’s requirements for seeing detail in the fine features. Projects that involve mapping the bare-earth ground surface for general engineering, flood risk or detailing the vegetated canopy may only require point densities of two points per square meter (ppsm), Meade says, explaining this as, on average, two 3D points captured from LiDAR on the ground in each square measuring 1 meter on each side. Projects like mapping above-ground electric lines may need considerably higher densities like 20 ppsm to ensure accurate capture of all line detail. Transportation projects that require specialized pavement analysis or urban projects that need to capture fine features accurately, wall irregularities or structural or HVAC details may require point densities ranging from 2,000 ppsm to 10,000 ppsm or more to achieve their goals, Meade says.
“The ability to collect highly accurate 3D models in a fast and cost-efficient way is the key advantage of LiDAR. Moreover, the application of the technology can be fine-tuned to each project’s requirements for accuracy and density that will provide the answers our clients are looking for at a cost they can afford.”
Photo Courtesy Quantum Spatial
LiDAR Modes
On a more specific and technical level, LiDAR is also categorized into different modes, tied to how the lasers and sensors operate to acquire data. Linear mode LiDAR, the most conventional form, sends out individual pulses of light and measures the incoming returns to determine position, elevation and intensity of each return. Meade says the sending out of individual pulses of light means more than one pulse can be in the air at any one point in time.
In addition to linear mode LiDAR, there are two newer forms, Geiger-mode and single photon, which, like linear, both use a laser to measure distance from the sensor to the ground and back. However, Meade says the way they send and receive the returned energy is different from linear mode LiDAR, and they collect many returns in the form of an array from each laser pulse. “The technology leaders, Harris Corporation and Sigma Space, believe they can fly higher and faster above ground than is typical for linear mode. The major promise is Geiger-mode and single photon generally allow collection of much larger areas at even higher densities within a single flight mission. Both are certainly interesting technologies that are still early in adoption,” Meade says.
The effective number of measured returns with linear LiDAR is a fraction of the measured returns from Geiger-mode or single photon. But, Meade points out, linear mode can be more effective in penetrating vegetation, allowing end users to see the natural ground below trees and brush, or having less noise in the acquired elevation surface. Riegl says it offers a form of LiDAR different from linear, Geiger-mode and single photon.
“The use of the term ‘linear’ relates to an older analog signal processing version of LiDAR, not the advanced digital regime of Riegl Waveform-LiDAR,” Riegl says. “The proprietary Riegl Waveform-LiDAR varies significantly from the so-called ‘linear’ LiDAR.”
The technical difference between linear and Riegl’s Waveform is tied to how signals from the receiving element, the photodetector, are analyzed to derive precise, highly-accurate and attribute-rich LiDAR data. Waveform LiDAR is extremely precise, Riegl says, because it provides high-ranging accuracy, low noise, high multi-target resolution, pulse shape information, time stamping and provides the basis for radiometric calibration on every point.
Waveform-LiDAR is designed to provide superior results through vegetation and leads to data where poles, wires and meshes can be seen. It provides near-real-time availability of processed point clouds, useful for applications such as rapid response or rescue operations. Waveform-LiDAR technology is incorporated into the Riegl VQ-1560i system, which can collect 450 square kilometers per hour of data at 8 ppsm.
Riegl says Geiger-mode and single photon LiDAR have the potential to collect more points, but at a loss of intensity and precision. “The dramatic increase in points with Geiger-mode or single photon LiDAR seems an overwhelming positive. However, the lack of precision and attribute information takes the data users back in time to an inferior result. All of us want our surveys and maps more precise, not less precise.”
At this stage, the newer LiDAR modes — Geiger-mode and single photon — don’t appear to pose an immediate threat to the relevance or existence of linear, according to both Meade and Riegl. However, that could change as limitations are worked out and/or if the newer modes prove more cost effective in dollars per square mile than linear LiDAR for wide-area projects, Meade says.
Photo Courtesy Quantum Spatial
LiDAR and Drones
Another area of innovation with respect to LiDAR is the advent of drones. “We are seeing an increase in activity now for LiDAR capture from drone platforms and there is a lot of promise in LiDAR for small project areas,” Meade says.
However, he says there are still a number of issues that need to be resolved from both an economic and performance perspective. While the Federal Aviation Administration’s (FAA) rules for commercial drone operations have expanded drastically, Meade says he finds that it is still often more economical to utilize manned flights in populated areas or for applications such as corridors that require beyond-line-of-sight flights.
He has noticed that the potential applications for LiDAR technology in the drone market worldwide have motivated many LiDAR manufactures to develop smaller sensor footprints at decreased weight and with lower power consumption. “Over the past few years we have seen the introduction of commercial LiDAR sensors that are specifically designed for small unmanned aerial systems (sUAS) resulting in a class of sensor that fills a niche between mobile mapping and traditional low-altitude airborne LiDAR.”
A key actor in advancing LiDAR technology for improved integration with drones is Velodyne LiDAR, which released the Puck Lite, its lightest sensor ever, in 2016. The compact LiDAR sensor has identical performance to Velodyne’s VLP-16 sensor, but weighs 590 grams as opposed to 830 grams.
“While LiDAR adoption in the drone space has shown considerable growth — five times year over year — the customer base remains fragmented with many regional integrators and start-ups carrying out the integration — mechanical, electrical and software — in-house,” says Harris Wang who works with market development and eco-system partnership cultivation at Velodyne LiDAR. As the market continues to scale, Wang says it would make sense for drone OEMs to provide factory-integrated systems so service companies can buy a ready-to-fly system. “Recent DJI integration of Velodyne Puck Lite is a sign of the industry shifting toward this model,” he says.
Photo Courtesy Quantum Spatial
LiDAR vs. Photogrammetry
Returning to fundamentals, a very common and basic question geospatial professionals pose is, “What is the difference between LiDAR and photogrammetry?” After all, both geospatial methods help generate 3D models of the environment. Wang offers up the table above, covering general differences.
Meade explains photogrammetry as the use of one or more overlapping sets of photographs that cover a project area to capture data. The overlapping images, known as stereo pairs, allow highly trained professionals to view the ground and built environment in 3D, and use specialty software to collect, one-by-one, the visible features to generate a detailed project model. Meade says many photogrammetry projects call for features like elevation mass points to model the natural ground surface; the limits of hydro features such as streams, ponds, lakes and the shoreline; building outlines; roads; bridges; and vegetation limits as a minimum.
“But consider this,” he says. “It can take a trained mapping professional 1.5 man years to collect 1 million elevation points on the ground through photogrammetry. Today’s LiDAR sensors collect the same number of points in one second of flight.”
In terms of acquisition, while imagery collection is passive, LiDAR uses an active sensor. With LiDAR, the sensor illuminates the ground with the laser so flights can be scheduled both day and night. This is drastically different from passive imagery acquisition, which is only optimal at midday, when the sun is at its highest elevation.
| Attribute | LiDAR | Photogrammetry |
| Ideal Application | Vegetation, complex or layered surfaces, small cross-section (power lines) and sharp edges (buildings/bridges) | Open area with smooth/clean, visually distinct objects |
| Limitation | Cost, no RGB color | Lighting dependent, vertical position accuracy |
| Data Processing | Real-time processing during flight | Post-processing due to data size |
| Integration Effort | Medium to High: IMU, RTK GPS integration for improved absolute accuracy | Low: Standard hardware with different camera configurations |
What’s Next for LiDAR?
In considering the most significant recent advancements with respect to LiDAR, Meade says the rapid change in the technology overall has been impressive and that innovative applications like topo-bathymetric LiDAR have opened many new markets. “Here, a visible green laser is used in the sensor design — as compared to non-visible, near-infrared energy in lasers built to measure the ground — that has the ability to penetrate water and provide returns on the sea floor, beds of rivers, or ground below ponds and lakes.
He says clients are demanding higher point densities in their datasets. Since storage needs are so strongly correlated to point density, Meade says both storage and computing power must not be overlooked. “Data storage technology has certainly advanced over the last few years and the cost per terabyte of storage has decreased, while the reliability of the storage and speed for accessing the data has increased. Within Quantum Spatial, we have 5 petabytes (5,000 terabytes) of active data storage available for our LiDAR processing needs. That is significant.”
With big LiDAR data here to stay, Meade expects to continue to see the demand from clients for more, with respect to density and accuracy of the data. Four years ago, he thought of high-density LiDAR data as 8 ppsm. Today, he tends to think of 20 ppsm as fitting with medium to high density. That creates challenges in storing the data and making it rapidly accessible to users, both during processing and during client use after project delivery.
All in all, Meade considers LiDAR an exciting and dynamic technology. He expects to continue to see improvements in acquiring data, processing the data into an accurate point cloud, and extracting information from the data to provide the answers clients need. With these improvements he expects continued cost efficiency improvements. “The usefulness of the technology is only limited by our imagination and creativity in applying it,” Meade says.
Geospatial professionals should know that the need to adopt LiDAR depends on the business practice and specialty of each firm or individual. Meade says most wide-area elevation projects today rely solely on LiDAR technology to develop the needed point clouds and only highly specialized projects or those with very limited area use other techniques like photogrammetry or ground-based kinematic GNSS applications to develop high-density point clouds. “While a new entrant [to LiDAR as a service offering] would have some catchup to do to the overall market, it is never too late to start.”
Original post seen here, written by Valerie King at POB.