XGRIDS Customer Success Story: Stantec
E38 Survey Solutions first heard about XGRIDS from another LiDAR manufacturer back in October. We were so impressed by everything we heard and the research we did, that in just 7 days we signed on as resellers of XGRIDS and placed our first stocking order. Maybe it was a bit of an impulse purchase, but the pace of demos/rental/sales of XGRIDS has justified the quick trigger.
The buzz and excitement surrounding XGRIDS hasn’t let up a bit since then. But, is it warranted? Is this actually a useful tool or just a cool new gadget that doesn’t really get regular use? Back in February, Stantec purchased their first XGRIDS K1. We sat down with their Reality Capture Technology Manager, Joshua Sexton to find out what he thinks of the K1.
Joshua has been in the Field Data Acquisition space for decades, and has operated just about every piece of positioning equipment; tape measures, total stations, RTK GNSS, drones, aerial LiDAR, SLAM scanners, etc. So, what would it take for Joshua to be able to put the K1 to work regularly? First of all, Stantec has tools that cover XGRIDS capabilities, but not necessarily tools priced low enough that they can scale to the desired level. Beyond scalability, it has to be fit for purpose, dependable, and repeatable.
Fit for Purpose
Stantec seeks both 3D Gaussian Splat outputs and point clouds. The Gaussians must be suitable for applications like virtual tours of restricted areas and virtual training sessions with live site walkthroughs.
Projects often do not require millimeter-accurate point clouds like those generated by terrestrial scanners. Usually, 3 cm or better accuracy is sufficient for RTK-denied uses such as underground or indoor areas. Tests showed that maintaining this 3 cm accuracy in RTK-denied areas is possible with proper control.
It isn’t all about RTK denied areas though. Periodically Stantec will perform clearance evaluations for transportation clients. In the past they have done this with terrestrial scanners. A lot of the clearances don’t change over time, so they have some trusted historical data. On a recent project they used the XGRIDS K1 and scanned obstructions over an 80 mile rail corridor, running the K1 with RTK corrections. They were really pleased with how everything came out. While they didn’t cross check every data point, spot checks consistently came in around 3-5cm, a level of accuracy acceptable for the effort.
Dependability
In the early days of drone photogrammetry, it wasn’t always easy. There were a lot of drone companies that came an went. For a considerable amount of time the most reliable mapping drone out there seemed to be the eBee. And the expected (crash) landing radius on an eBee covers about an entire football field. Then, there is a whole slew of other problems of cameras not triggering or mismatched geotags and all sorts of issues. Then, came DJI. Even they had fly aways early on. Now you can train someone on how to operate a DJI drone in about 10 minutes and have complete confidence that it is going to land in one piece with the data you need.
We talk about SLAM scanners in a similar vein. There are some sort of DIY products out there that do one thing well, but are retrofitted to SLAM and just disappoint. And there are other stronger solutions, but still lack features like RTK, and even then, the price is prohibitive to scaling.
Part of the excitement with XGRIDS is that it is a SLAM scanner that actually works. We’re talking about a tool with reliability far closer on the spectrum to DJI than many other similarly priced scanners on the market. And it’s a SLAM scanner with a really nice feature set. It’s a SLAM scanner with integrated RTK. It’s SLAM solution with a 3D Gaussian Splat software built from the ground up for their hardware specifically. You can pull the 3DGS right into Unreal, or you can use Revit with the new LCC plugin, and even process 3DGS from your drone and combine them all in Lixel CyberColor. And annual fees? Yes, on the 3DGS software Lixel CyberColor, but if you just need the point cloud processing, you receive permanent licenses of LixelStudio.
On the question of whether it is dependable, Joshua put it well when he said, “I pushed the K1 to do a lot of dumb stuff, but it never failed.”
Repeatability
Really the question here is, can we ship this to a subject matter expert (SME) with a quick how-to video, confident they can collect acceptable quality data for their intended purpose? It’s a really important question for Stantec that makes a huge difference. They are a firm built of SMEs first and foremost. Their teams are looking for the right tool for their job, rather than a LiDAR firm looking for the right job for their tool.
So, it was really attractive to Stantec that XGRIDS seemed to have their own easy button for SLAM scanning and 3DGS.
Conclusion
At the end of the day, what does Stantec think of the XGRIDS K1?
Data capture has been easy with the LixelGO application. When fit for purpose, they can collect data at a 2-5cm accuracy in RTK denied areas and outdoors with RTK corrections. Pulling in the 3DGS into Unreal Engine has worked out really well.
It’s keeping their teams off ladders, out of buckets, and away from asbestos covered piping.
It’s capturing so much more data than a tape measure, so that they have less return trips because of additional measurements needed later.
The 3DGS is detailed enough for a lot of their work, but not good enough to read labels with the K1 alone. XGRIDS has released a new HD enhancement feature which they are hopeful will allow them to do more with the K1.
It is fit-for-purpose for a lot of use cases, dependable, repeatable, and scalable.