In the last several months, discussions concerning Artificial Intelligence seem to be everywhere. One of the key reasons AI has garnered so much attention of late, is that by some measures, companies like OpenAI (makers of ChatGPT) and others are on the verge of resolving, or least greatly mitigating, what’s known in computational, biological and mechanical sciences, as the speed-accuracy tradeoff, or SAT.
Modern AI is now fast enough and accurate enough (though still far from perfect) to approximate natural conversation with both speed and accuracy that mimics human-like exchanges with a computer’s speed of data retrieval. This same balancing act between speed and accuracy plays itself out in 3D laser scanning and its related offshoots like panoramic 360° photography and photogrammetry.
In traditional 3D reality capture high-accuracy scans can take minutes to capture and much longer than that to process the information and register the point-cloud data. Accuracy at the loss of speed. Which for any company looking to complete large-volume scans can mean a potential loss of business. A panoramic camera, however, can capture an image with great speed but the corresponding data is far less detailed than a 3D point cloud. Speed at the expense of accuracy.
But if AI research suggests that the end of the speed-accuracy tradeoff is in sight, what is the next digital domino to fall? Not one to be left in second place, the 3D laser scanning industry, which includes 3D visualizations for architecture, engineering, construction and operations and maintenance, and public safety pre-incident planning applications, is poised to claim victory here too.
A World Waiting to be Scanned
Today, those in the business of as-built documentation, are also continually looking for faster, better, easier ways to get the job done. They are looking to simplify tasks and streamline workflows — without sacrificing quality for speed and speed for quality.
Why are they looking for this “holy grail” of speed and accuracy? Because the entire professional measurement services industry has a lot of work on their hands. And lots of time, money and human and technological resources are at stake if BIM models aren’t accurate or if image capture and processing speeds are too sluggish. Across the globe collectively the industry is scanning the interior and exterior of large structures; they are mapping and measuring industrial facilities like power plants, they are assisting police and law enforcement with their pre-incident planning, they are aiding specialty and general contractors in their engineering projects, and they are empowering facility managers/owners with the real-time facility management insights they need to track building operations and to streamline and eliminate any inefficiencies they uncover.
As might be anticipated, the quest to resolve the speed-accuracy tradeoff ultimately took two distinct paths — mobile scanning products and static scanning tripod-mounted 3D laser scanners.
But instead of truly ending the SAT “battle” all the development really did was carve out two spheres of competing influence. Repeated studies like one conducted by the University of Cambridge — which included a FARO Focus Laser Scanner — and shared at the 2021 European Conference on Computing in Construction on the island of Rhodes, Greece, confirms that mobile scanning, while fast, is not as accurate as a static scan.
A section of the paper’s conclusion reads:
“Our experiment showed that the density of the scanned targets measured in points per second decreases exponentially along with the growing distance to the target and that the mobile scanners outperform the static mapping devices in this regard. The trend is the opposite when it comes to accuracy. The static scanner produces scans that are at least 20 times less noisy than those by mobile devices. While the accuracy specifications for such use cases as measured building surveys, topographic surveys and low accuracy setting out are met by all the devices almost at the whole range up to 40 meters, the more demanding use cases such as engineering surveying could only be satisfied by the static scanner.”
Hybrid Computing 2.0
If humans and computers were what counted as “hybrid” in the 1960s at NASA (Human “computers,” were hired to crunch critical launch and re-entry data in part because the space agency’s mainframe, the IBM 7090, while faster than humans, wasn’t as accurate as their genius) for 3D reality capture hybrid is about to mean a union of two types of digital formats, uniting the best qualities of mobile scanning with the best qualities of static reality capture.
This union will be in the form of a new technology and a patent-pending term. With this hybrid technology, an algorithmic merging of the data contained in a panoramic image, combined with a static 3D point cloud, a lower resolution 3D laser scan is augmented with data captured from a 360° camera. The result is that for the first time, speed and accuracy are no longer sacrificed for each other.
Hybridization of mobile scanning and static scanning with minimal loss in accuracy and speed is the next-step in 3D scanning solutions. Consider the implications of a technology that is up to 50% faster than traditional laser scanning workflows. And one that provides the best combination of onsite productivity, predictable accuracy and state-of-the-art visual clarity at a highly affordable price.
The benefits of such a merging should be clear:
- A scan that is reduced in accuracy designed to be significantly faster than traditional scans.
- 360° images that augment less dense data sets to ensure project completion with accelerated workflows
- The saving of entire days per week, enabling the taking of additional scans that would likely not have been taken without the added speed and time savings, ensuring more complete and comprehensive projects.
- Colorized images that allow users to see object edges and corners more clearly and identify defined geometries with greater ease, leading to better analysis and execution in other software.
While the technological innovations that got the 3D laser scanning industry where it is today are evolutionary in nature — the first laser scanning devices were developed in the 1960s and used lights, cameras and projectors to scan objects but were time consuming and prone to error — a merging of static scan data and mobile 360° images with no loss of speed and accuracy is revolutionary.
A reality capture solution that’s up to 50% faster than traditional scanning with virtually no loss in data quality means that industries reliant on large-scale scanning, from architecture, engineering, construction and operations and maintenance to public safety, will now have a fast, accurate and reliable way to take thousands of scans and use that data to drive business results or achieve situational results, like the fast scanning of entrance and egress points to a large building or the location of obstacles for access. Considering that there are approximately 100 billion buildings in the world and that a sizeable percentage of that number includes large-scale construction, the opportunity to accelerate the conversion of the physical world into the digital world, has never been greater.
It’s as if a NASA human computer was given access to Frontier, currently the world’s fastest computer, with total confidence that its one quintillion floating point operations per second, was 100% accurate.
New Frontiers
Whether it’s the physical frontier, the biological frontier, the mechanical frontier or the computational/digital frontier whose limits are pushed by today’s fastest computers and ever smarter algorithms, what has been true since the first cities emerged from the Fertile Crescent of the Eastern Mediterranean thousands of years ago remains true today. As long as humans populate this world, large, complex structures, will be built, requiring fast, accurate assembly and long-term project oversight. And for public safety, the need to mitigate worst-case scenarios before they occur will be ever-present.
With these two realties a given, 3D laser scanning will continue to advance. Not only will speed and accuracy increase, but the equipment’s size, weight and bulk will continue to decrease. While app-based LiDAR scanning is beyond the scope of this article, advances in generalist technology, no bigger than a smartphone or tablet is where the future ultimately lies.
Combined with the promise that quantum computing offers — the world’s first universal quantum computer with more than 1,000 qubits is slated for rollout by IBM later this year — the new frontier is likely to push well beyond the speed-accuracy tradeoff in the decades to come.
But for now, hybrid reality capture is the technological breakthrough the 3D laser scanning industry is just beginning to capitalize on. The speed-accuracy tradeoff, long the essential pain point for a variety of disciplines across time, is increasingly a challenge looking for a solution that’s about to be solved. Acclaimed British science fiction writer Arthur C. Clarke may be known for his often-quoted observation that: “Any sufficiently advanced technology is indistinguishable from magic.”
But while true, Clarke’s quote says nothing about how exciting the magic can truly be — if you’re one of the visionaries helping bring that magic to life.
About the Author:
Oliver BürklerOliver Bürkler is the Director of Laser Scanning, FARO Technologies, Inc. He earned his masters in precision engineering and business administration & engineering at the University of Applied Science in Munich, Germany. As part of the FARO 3D laser scanner product manager team, Bürkler is focused on the development of FARO’s 3D documentation hardware and related new innovations.
Excited about this new technology and how it can benefit you and your workflows?