GEO Business 2019 Shows Equal Doses of Caution and Excitement Toward SLAM

The GEO Business 2019 show, which took place in London in late May 2019, showed a receptiveness of the surveying industry to new software and forms of handling the fast-growing amount of data generated by surveys. Not by accident, the event took place in tandem with the GeoDATA Forum, which included a number of seminars on managing point clouds, data privacy, and analyzing data from Geographic Information Systems (GIS). While the software and analytics side of the show seemed focused and purposeful, the hardware Original Equipment Manufacturer (OEM) side was less so. Some businesses bet on the traditional total stations for surveying, with added benefits like theft protection, while others believed surveying robots for highways could render total stations obsolete in some cases. Some believed that inspection should be performed with accurate Light Detection and Ranging (LiDAR) point clouds, while others were in favor of the power of 360° Red, Green, and Blue (RGB) cameras. The current multifaceted polarization of the industry is better encapsulated by the acronym that managed to become both a buzzword and a dirty word at the event: SLAM.

Simultaneous Localization and Mapping (SLAM) is a technique that allows a device to map its environment while it roams through it without requiring preexisting maps or an external location aid. SLAM is particularly relevant when a device does not have access to Global Navigation Satellite System (GNSS) positioning. In a survey, a LiDAR scanner with an Inertial Measurement Unit (IMU) will usually be carried around a venue by a person as it generates a point cloud for later analysis. The benefit that SLAM provides surveyors in construction or for Building Information Modeling (BIM) companies is that these digital twins can be generated in a fraction of the time and with a fraction of the effort and manpower that common surveys require. For example, an entire 13-story building can be mapped by one person pushing a cart around or wearing a special backpack in less than two hours, whereas traditional surveying techniques could take up to a few days. This possibility is why SLAM became a buzzword.

The surveying industry, however, is fairly accuracy-oriented. When performing construction inspections, surveyors and engineers seek to build and analyze models that are accurate down to the millimeter. In industrial hardware inspection, digital twins of machines must also have mm-level accuracy to provide visibility of the smallest damages or faults. Currently, such accuracy cannot be reached by SLAM algorithms. At best, a LiDAR-based SLAM surveying scanner can reach an accuracy of 2 to 3 cm. Hence, lest certain companies be associated with low accuracy devices, they decided to distance themselves from SLAM and became wary of using the acronym to advertise their products. This limitation is why SLAM became a dirty word.

The result of this was a show in which only one company, GeoSLAM, touted and advertised its SLAM capabilities unapologetically, with its new ZEB-REVO scanner. Other companies, like NavVis, were shyer with their offerings, despite boasting an impressive SLAM solution with high-accuracy mapping that does not require loop closures and can correct drifts that tend to happen in long hallways. There were many companies that indeed made use of SLAM algorithms at GEO Business, but one had to push them a bit before they used the S-word. FARO Technologies presented ScanPlan, a SLAM-enabled handheld laser scanner for creating 2D floor plans. The scanner can make creating floor plans and calculating flat areas for the real estate market cheaper and more convenient. For this use case, the accuracy of SLAM is enough. Leica Geosystems presented its Pegasus:Backpack, which maps the environment around the bearer using LiDAR and high-precision GNSS. If a GNSS signal cannot be obtained, the device uses SLAM for mapping. In GEO Business 2019 there were plenty of SLAM offerings, they were just not immediately obvious.

The accuracy limitations of SLAM mean that this technology will only have access to a fraction of the surveying market until it can improve its accuracy by an order of magnitude. Today, it has enough accuracy for use cases like stockpile volumetric surveys, BIM, and floor plan generation. As shown by NavVis’ solution, however, software innovations can lead the way to better accuracy. There are relatively few companies specifically working on improving SLAM algorithms today, and therefore a market opportunity for that becomes evident. The wave of innovations across the machine learning landscape and sensor fusion could also be a driver for better SLAM accuracy and software, but it will require some companies in the surveying ecosystem to relinquish their prejudice.

Finally, better SLAM algorithms can not only unlock most of the surveying market, but also promote disruptions in other verticals, like Autonomous Vehicles (AVs) and Augmented/Mixed Reality (AR/MR) if such algorithms can be made lighter. SLAM can also enable autonomous robotic inspections in hard-to-reach or hazardous places. Below the waterline, sonar-based SLAM will likely become important for bathymetric surveys and marine cable inspections with the development of Autonomous Underwater Vehicles (AUVs). Such will be the innovations presented at future GEO Business shows.