Technologies developed for autonomy and consumer applications are opening up new user segments and rewriting conventional workflows
A construction inspector sweeps their phone over an open trench creating a 3D as-built of the utilities within. This is just one example of low-cost, rapid, simplified reality capture is challenging conventional wisdom on how reality capture should be executed. In addition to this new pervasiveness of consumer and pro-sumer lidar tech, more conventional lidar systems have also seen dramatic improvements.
Yes, there is always the caveat that diligence and skill combined with conventional large-format scanners cannot be beat in terms of precision, accuracy, range, and detail for surveying and other AEC applications. However, value propositions are being challenged when a wider range of applications—with fit-for-purpose precision—come into play. This opens up 3D for many more applications that may have been somewhat impractical, and/or too costly to consider doing with conventional scanners.
Phone and Tablet Lidar Sensors
It has only been just over two years since the release of the iPhone 12 with a lidar sensor, and developers wasted no time in seeing how far they could stretch the technology for a wide range of applications. There would be too many to list, but we have a few examples for real-world AEC applications.
Data exports for two scanning methods compared in CloudCompare. The rooftop telecom asset lower scan (created by a $20,000+ scanner) was done with standard setups, and the iPad capture (upper) was done walking around the roof, keeping less than 5 meters from key assets. The latter was done in a quarter of the time of the conventional scan. While both had some rough edges, the iPad capture was more complete for key assets.
Improving the efficiency of telecommunication infrastructure mapping, like cellular sites, has been a focus for Virtual Technology Simplified (VTS). VTS determined that capturing the antennas and equipment on cellular towers was most efficiently done with a UAS and photogrammetry, and they’d developed tools and workflows for doing these. But non-tower cell sites proved to be a challenge.
“Twenty-five percent to 30 percent of cell sites in the U.S. are on rooftops,” said John Chwalibog, VTS founder and CEO. “Even more in Europe and parts of Asia. In many urban areas, capturing those with a drone is prohibited, and might not be the best option to capture what is needed. Plus, a tech would need to do a site walkthrough anyhow, to capture additional information.” With conventional scanners that cost upward of $20,000, a single site scan could add up to $10,000. “Telecoms are used to a $2,000 drone being used to rapidly map cell towers.” Said Chwalibog. “We started looking at consumer level devices, to keep the costs down and speed up operations, but also to fully script the process so that just about anyone could perform the site mapping reliably with minimal training.”
To ensure that their envisioned solution could meet precision requirements, VTS did a rigorous study, contrasting conventional scanners with iPhone and iPad lidar/image sensors. (Download the white paper here: bit.ly/3VNvKZx.) The iPhone tested includes Apple’s A15 bionic chip and 12MP TrueDepth Pro camera system. The iPad tested included Apple’s M1 chip and12MP TrueDepth Pro camera system. Point clouds were produced using both lidar and imaging, using EveryPoint Lidar Fusion. Scanners tested included a conventional mid-range scanner, a scanning total station, and a mobile cart system, costing between $35,000 to $75,000.
Resultant point cloud from two scans of a rooftop generator enclosure: a conventional terrestrial laser scanner (left) and an iPad (right) with the flash lidar and images processed together. While the pad capture produced finer detail, the range is very limited, though for this type of roof capture it was sufficient.
Relative results with the phone/tablet for planar, spherical, and floor flatness were not significantly degraded from those of the scanner and scanning total station, and comparable to the mobile cart (if not better in certain situations). Position was a bit worse though for the application, relative precision and detail are more important. The combined lidar/image processing yields fine detail. But a caveat, these great results for the phone/tablet lidar are for ranges under five meters. This is just fine for the intended workflow: walking the roof to capture the cell antennas and associated hardware. Range for these sensors, and many other low-cost sensors is a handicap, yet they are perfectly fine, if not superior in some respects, for short range, fit-for-purpose results.
The real benefits come from not just the low hardware costs ($1,500 – $1,800) but in the workflow improvements. VTS has developed 360Capture, an app that connects to the cloud to accommodate processing and managing field tech operations. Chwalibog said they did a lot of research, looking at the workflow, and ran many field tech scenarios. He jokes that they even looked at one for a hypothetical hungover tech that really didn’t want to be at work that day. They wanted to know how they could make any tech more efficient by identifying any step that might disrupt the tech, like not having a good on-site contact for roof access. VTS feeds everything the tech needs to the app, including cell carrier requirements for each site capture.
“The whole idea is all of this information about the job is contained in a single location for the technician to make it really easy for them to be prepared and start to go to work and do what they need to do,” said Chwalibog.
Time saved at each site can be substantial. Prior to developing the phone/tablet lidar solution, their go-to was light-weight $20,000 scanners, or some image-only systems. These were cumbersome and time consuming. The capture aspect of an average sized roof capture would take between 40 minutes and several hours. With the new solution they could often do the same in seven to 15 minutes. Standard outputs, like e57 point clouds are easily ingested into utilities applications like Bentley Systems Open Tower.
Another solution using phone/tablet lidar that is growing in popularity for utility as-built is Pix4D’s viDoc. It is an RTK rover that can be used on a pole or handheld (there is a downward facing laser that can be used, for instance, for ground control points). You attach a supported laser-equipped phone/tablet, sweep over the area of interest at short ranges, and create a point cloud that is spatially registered.
Phone and pad scanning has hit mainstream AEC applications. This handheld device (called the viDoc) is a full constellation RTK rover that you attach the phone/pad to, providing orientation and positions for the scans/images. In this example, the assets in the open trench are captured rapidly as an as-built before being buried.
These are just a few. There are many apps for forming point clouds from phone/tablet lidar/imaging, even some open-source options for developers and DIY. A key takeaway from this boom in leveraging tech that emerges from the consumer sector is that it opens up many more opportunities for many more people and enables “crew sourcing” of 3D reality capture. “We really wanted to be part of the ecosystem that’s trying to create scalability and all of this, how we get more people doing more 3D capture,” said Nathan Pugh, interim CTO for VTS. “3D for the masses is a reality.”
Solid State Lidar
There are many types of solid state lidar, like a form of flash lidar used in the iPhone/iPad. There are some types that include micro electromechanical systems (MEMS), but like the standard spinning-mirror lidar scanners that have been standard for so long, there are still moving parts. What many call “true solid state lidar” has no moving parts. They employ methods like beam forming to scan a wide field of view. Such sensors are gaining in popularity for autonomous vehicle applications. They can be made smaller, can be more resistant to vibration in some instances, and can be much less expensive than traditional mechanical lidar.
We are now seeing solid state lidar in surveying, mapping, and other AEC applications. An example is XenotroniX, that has integrated these sensors into a mobile mapping system, with a further focus on roadway/pavement management, but suited for other corridor mapping uses.
“We wanted to get the motor out of lidar,” said Karsten Bronowski, sales and business development manager for XenomatiX. “When you have moving parts, you have wear and tear, the effects of vibration, problems with long-term reliability, and with controlling temperature. With true solid state lidar you can eliminate these issues.” Bronowski said their systems feature a CMOS-based detector generating high-density point clouds in all-weather conditions, a multi-beam laser projector generating a high-resolution grid of points.
Their dual lidar sensor system further utilizes triangulation but gets its orientation and positioning from additional components, including GNSS and IMU. The system that Bronowski demonstrated at Intergeo in October 2022, has a Septentrio AsteRx-U3 GNSS/IMU unit supporting dual antennas for heading, and a separate camera. They have a few additional proprietary components to ensure sustained positioning in GNSS denied areas. Bronowski said the system can be mounted on a vehicle facing forward for preview applications, like for heavy transportation users as a reference for managing active suspensions. Or it can be mounted rear-facing for high-detail pavement scanning, for pavement management applications.
There’s a lot of difference of opinion on solid state lidar, as there would be with any emerging, potentially disruptive technology. One thing is for sure, the development and application of this tech points to broader adoption. I would not be surprised to see low-cost, solid state static scanners on the market within a year or two—or added to GNSS rovers and total stations?
Commoditization of Lidar Sensors
Small pucks and box-style scanning sensors, ranging in price and capabilities, have been available for years, and the list of manufacturers continues to grow. Many of the mobile, UAS, and backpack mapping systems feature these OEM sensors, like those from Velodyne and Livox. Again, development for autonomy, robotics, and industrial applications has been a key driver. Some of these sensors are so inexpensive that even hobbyists are able to integrate lidar, especially with so many software choices, including open source.
Small, powerful, and affordable modular lidar sensors can be easily integrated into systems for numerous applications. An example is the Ap-Lidar-One, a lightweight yet powerful SLAM scanner with built-in processing.
This boom in OEM lidar sensors has spawned a new wave of development of third-party enhancements. An example we spotted at Intergeo 2022 is the Ap-Lidar-One from Aerial Precision. It is a lightweight, yet powerful SLAM scanner with built-in processing. You can export industry standard format files on a data card upon completion of scanning. It is based on an Ouster OS1: 1.3M points per second with a +/5 cm accuracy at 120 m. They offer several mounting options: standard tripod thread, and adapters for most of the popular UAV platforms.
Products like this, and there are many out there, open up a lot of opportunities for sensor integration. It is getting to the point that you can stick lidar on just about anything that moves, scanning everything it sees. Visions of digital twins, with “continuous representation of reality” would need to feature such ubiquitous lidar.
A Continuum of Development
The traditional big players in the scanner market have not stood still. We have seen vendors fill-out their scanner portfolios, especially the relatively recent addition of mid-range systems like Leica’s RTC360, and Trimble’s X7. The RTC360 and X7 feature auto registration of point clouds (using different methods, but both with excellent results). The X7 also features self-calibration, that could mean the end of need for sending it in yearly. With such features, plus automated fine-leveling, some in the industry say they are evolving into “cobots” (bit.ly/3XTudCZ).
A wave of mid-range scanners, like the Leica RTC360, Trimble X7, this Riegl VZ-600i, and others could take over much of the work that had traditionally been done with legacy large-format scanners. This trend also includes scanning total stations that users are leveraging for an increasing number of applications.
And Riegl, renowned for making some of the highest quality sensors, especially the high-ticket units, has added a new mid-range terrestrial scanner, the VZ-600i. It differs from most mechanical scanners in that it has two spinning mirrors in the enclosure. This provides the speed and density of some of the larger systems.
We did notice that it does not do “full dome” scans. But as David Foster, managing director of Riegl UK, pointed out during a demo at Intergeo, “Subsequent setups of the scanner would pick up features above anyhow.” It also featured something else we are seeing on many new scanners: very large onboard screens with very well-thought-out menus for simplified operation.
Mobile mapping has come a long way. Backpack units have evolved from some rather odd contraptions, to fully thought-out systems that are more precise and simpler to use. There have been sophisticated backpack systems for many years, like the Pegasus, but we see more of them now, like the NavVis VLX and Green Valley LiBackpack.
A bigger leap forward has been in vehicle-mounted systems. In recent years, such systems have been able to break the centimeter barrier, though with a lot of control points. The new Leica RTK appears to be able to do this with little or no control. It may be among the most sophisticated mobile mappings systems so far, a culmination of two decades of development. In addition to the high-end scanner head (or two-head version), the 360-degree camera is paired with fisheyes on a post, angled at about 45 degrees from the direction of travel, to help eliminate stitching issues. There is full-constellation GNSS built into the body of the unit, and a tactical grade IMU. There are optional butterfly cameras you can add to either or both sides, but there are two other features that really stood out.
They’ve added a SLAM scanner to the front and rear of the unit, not necessarily to collect more points (though you could post-process it if you’d like), but to help refine the positioning. “SLAM addresses the problems encountered in GNSS-limited or GNSS-denied environments, like urban canyons, under a heavy tree canopy, or in tunnels,” said Alessandro Nuzzo, product manager for mobile mapping at Leica Geosystems. “When GNSS is interrupted for more than 60 seconds, the IMU starts to drift substantially. This SLAM integration can maintain high accuracy in those situations.”
“We learned a lot from our experience, and that of our customers, using our Pegasus: Backpack system, that was designed for indoor use. If you have a 360-degree field of view with the SLAM, you can better calculate the trajectory,” says Nuzzo. “Because a horizontal field of view works best for SLAM, we mounted them that way. If we had mounted them vertically, you would only see a profile for that spot; horizontal gives a wider field of view.” I’ve spoken to some early adopter firms, and the improved positioning is definitely working as stated.
With so many mobile mapping and imaging systems plying streets and highways on a continuous basis, privacy concerns have risen. There are new requirements, like the General Data Protection Regulation (GDPR) of the European Union. The TRK uses an AI-driven anonymisation algorithm that obscures certain features before anything is written to storage. This ensures compliance in that at no stage of the processing or data handling are there any stored images with faces and licence plates.
Quick Takes
Over the past year, surveying and engineering firms have noted that their use of the large terrestrial scanners has decreased, and they’ve found the newer mid-range scanners to be very versatile, as has the use of scanning total stations like the Leica MS60, Trimble SX10/12, Topcon GTL-1000, etc. Crews are finding that for many scanning jobs, they can use the total station, with power of the total station for establishing control, and with some systems providing instant deliverables. Certainly, they are slower than a dedicated scanner, but as I found in using one, while it was doing the scan and image sets, I took a GNSS rover to pick up additional points that the scanner would miss.
It is like UAS lidar has suddenly become the preferential payload for many UAS operators. It does provide much more flexibility for more applications than photogrammetry alone, but there were few affordable options in the past. While they are still quite expensive, some of the lidar systems designed for UAS are matching many of the specs of their terrestrial cousins. Like the Yellowscan Voyager that has a Riegl VUX120 scanner head, a range of over 700m, and yields up to 15 returns. The capabilities of lidar on UAS has improved so much that transportation projects, originally needing terrestrial campaign style scans to meet precision requirements, then mobile mapping (as that improved), are now turning to UAS lidar.
Andrei Gorb, segment manager for mobile mapping and UAV systems at CHCNAV, says he has seen an increase in UAS lidar use, with recent years of development seeing transportation project requirements being easily met, flying between 50m and 100m. Plus, many UAS can stay in the air longer, particularly the VTOL models.
If the makeup of the exhibits at Intergeo 2022 is any indication, the present boom in tech for surveying and mapping is in lidar and integrated systems. Five years ago, it seemed that half of the exbibits were drone platforms. There are far less of those now, but much more lidar, so many exhibits that we could not get to all of them in three days.
A reflection on the evolution of lidar for AEC: we just sent our old big-box scanner (the size of a dishwasher) to surplus. It was a miracle in its day. Now you can hold that power—and much more–in the palm of your hand. Coolness ahead.