FARO ACQUIRES GEOSLAM

To chat about anything else.
Tanguy Nédélec
I have made 70-80 posts
I have made 70-80 posts
Posts: 71
Joined: Fri Oct 23, 2015 3:15 pm
7
Full Name: NEDELEC Tanguy
Company Details: Prigent et Associes
Company Position Title: Surveyor - nooks and crannies explorer
Country: France
Linkedin Profile: No
Location: Dinan
Has thanked: 4 times
Been thanked: 11 times

Re: FARO ACQUIRES GEOSLAM

Post by Tanguy Nédélec »

pburrows145 wrote: Fri Sep 09, 2022 10:07 am Software update in REG360... check release notes.
...
Hope that helps.
It does, thanks! Thought it was a firmware update for some reason. Time to renew our licence!
User avatar
gsisman
V.I.P Member
V.I.P Member
Posts: 658
Joined: Fri Oct 07, 2016 1:51 pm
6
Full Name: Steve Long
Company Details: Montgomery County DOT _ MD
Company Position Title: Land Survey Supervisor
Country: United States
Skype Name: gsisman1
Linkedin Profile: Yes
Has thanked: 437 times
Been thanked: 97 times

Re: FARO ACQUIRES GEOSLAM

Post by gsisman »

smacl wrote: Tue Sep 06, 2022 9:48 pm
gsisman wrote: Tue Sep 06, 2022 7:28 pm
smacl wrote: Fri Sep 02, 2022 5:12 pm

Working with some client topo data collected with NavVis over the last few weeks and it is excellent, far better than anything I've seen from other SLAM devices. Not ideal for tall surveyors in tight spaces, but great outside of that.
It's fun watching Marek 6'-9" or 6'-11" do a demo. We had him in a parking garage and there was a lot of ducking.
If he were an instrumentman with tradition equip he'd need one of those 7' tripods!
Their data is clean.
I want to get another good dataset because our first one with a consultant still had some issues with the colorization outdoors
I think one of the other things to think about for topographic work is control. Problem with clean data is that is easily confused with accurate data, which is a regular issue with scanning. I do like how NavVis are strong on incorporating external control on the VLX and have been from the outset. While it adds to site time and cost, it does make it a proper surveyor's tool. Looking at the intensity data on a recent job, I think it is synthesised from the rgb, which isn't great but not surprising as intensity returns often seem to be a weak point from multi channel scan heads. BLK2GO is better here from the data I've seen. Also, definitely great job prospects for shorter surveyor's in the near future 😀
Yes, that is what I liked about it -the ease of acquiring an exterior control point was more like a traditional workflow. The 3500' walk scan we had through a hillside neighborhood just inside the beltway had probably close to 100' vertical elevation drop and we had RTK's and traversed control with our TCRP1201. When all was said and done the scan cloud was within0.1' vertically all over that site compared to RTK check lines we ran...(and that could have been the RTK vertical error)
VXGrid
V.I.P Member
V.I.P Member
Posts: 370
Joined: Fri Feb 24, 2017 10:47 am
5
Full Name: Martin Graner
Company Details: PointCab GmbH
Company Position Title: Research and Development
Country: Germany
Linkedin Profile: No
Has thanked: 121 times
Been thanked: 121 times
Contact:

Re: FARO ACQUIRES GEOSLAM

Post by VXGrid »

smacl wrote: Mon Sep 05, 2022 4:07 pm [...]

There are quite a few components that contribute to the quality of a SLAM derived point cloud. For the NavVis stuff the most obvious one is the excellent post processing software, but also the twin scan heads, four cameras, IMU, and more stable platform. To my mind it is much closer to the Leica Pegasus backpack in how it works and also the quality of the final result. The Pegasus backpack is also twin scanner, multiple cameras and has a profiling laser with excellent intensity results, but then also costs a whole lot more than the NavVis and is a bit dated at this point. I don't think currently available handheld SLAM devices are comparable and Faro would need to do quite a lot to have a comparable product.
What I am woundering is, if the scan head is a lot better with less noise, if the resulting cloud is better as well.
I mean if the single point accuracy is better, the positioning part should be better automatically.
The question is: Will the result be better than an improved algorithm, or can the results from NavVis only be reached by using two scan heads.

Looking forward to what FARO and GeoSLAM are going to manufacture in the future.
Once the static registration accuracy is on par with SLAM, we will see the market shift very fast I guess.
User avatar
Scott.Warren
V.I.P Member
V.I.P Member
Posts: 334
Joined: Tue Jan 30, 2018 3:14 pm
4
Full Name: Scott Warren
Company Details: WSP Canada
Company Position Title: Project Manager - Geomatics Engineer
Country: Canada
Linkedin Profile: Yes
Location: Dartmouth, Nova Scotia, Canada
Has thanked: 35 times
Been thanked: 95 times
Contact:

Re: FARO ACQUIRES GEOSLAM

Post by Scott.Warren »

I don't agree with an upgrade to one part guaranteeing it being a better overall solution.

If a total station is only accurate in angular measurements to 40", having a very accurate EDM doesn't really ensure the Traverse you run will be accurate. The summation of errors and the 40" accuracy (in this example) will detract more than the EDM gains will help. Accuracy is a summation of error sources, and a big source of error that isn't accounted for doesn't get removed from the equation.

I believe (I'm not 100% positive) the VLX also uses the images along with the 2xLiDAR sensors (could be mistaken) when processing. I believe NavVis has the edge because its more than just the traditional Slam (IMU + LiDAR Sensors) being used, like the BLK2Go 'GrandSlam' idea merging the 'live tachometry' derived from the images into the SLAM/trajectory mix as well.
VXGrid
V.I.P Member
V.I.P Member
Posts: 370
Joined: Fri Feb 24, 2017 10:47 am
5
Full Name: Martin Graner
Company Details: PointCab GmbH
Company Position Title: Research and Development
Country: Germany
Linkedin Profile: No
Has thanked: 121 times
Been thanked: 121 times
Contact:

Re: FARO ACQUIRES GEOSLAM

Post by VXGrid »

Scott.Warren wrote: Tue Sep 20, 2022 2:50 pm I don't agree with an upgrade to one part guaranteeing it being a better overall solution.

If a total station is only accurate in angular measurements to 40", having a very accurate EDM doesn't really ensure the Traverse you run will be accurate. The summation of errors and the 40" accuracy (in this example) will detract more than the EDM gains will help. Accuracy is a summation of error sources, and a big source of error that isn't accounted for doesn't get removed from the equation.

I believe (I'm not 100% positive) the VLX also uses the images along with the 2xLiDAR sensors (could be mistaken) when processing. I believe NavVis has the edge because its more than just the traditional Slam (IMU + LiDAR Sensors) being used, like the BLK2Go 'GrandSlam' idea merging the 'live tachometry' derived from the images into the SLAM/trajectory mix as well.
Scott, your are totally right with your opinion that there are always multiple adjusting screws to improve performance, and if one of the best performing parts is exchanged for a slightly better one, there won't be much or no improvement at all.

One way to improve mediocre hardware is on the software side, through intelligent algorithms, but this is always depending on certain assumptions (in which environment do I scan, can I integrate multiple sensors, how robust is this approach, ...)
Sure we need all components to work together, IMU, scan head, SLAM software to get a decent result.
If I slap together a 40k IMU with a 0.1mm accurate scanner head and my SLAM approach is bad, the result will be bad and who is going to buy it, if my device costs 3 million anyway.

As far as I extracted the information from GeoSLAM marketing material, their SLAM focuses on loop closures, so works mainly on the captured cloud, which gets improved with a better scan head, therefore I think the results will be better.
Sure in challenging environments with bad overlapp it will mainly use the IMU and then a better IMU will provide a better result, but exchanging the velodyne for one from FARO for your standard flat?
User avatar
smacl
Global Moderator
Global Moderator
Posts: 1073
Joined: Tue Jan 25, 2011 5:12 pm
11
Full Name: Shane MacLaughlin
Company Details: Atlas Computers Ltd
Company Position Title: Managing Director
Country: Ireland
Linkedin Profile: Yes
Location: Ireland
Has thanked: 457 times
Been thanked: 466 times
Contact:

Re: FARO ACQUIRES GEOSLAM

Post by smacl »

VXGrid wrote: Tue Sep 20, 2022 2:37 pmWhat I am woundering is, if the scan head is a lot better with less noise, if the resulting cloud is better as well.
I mean if the single point accuracy is better, the positioning part should be better automatically.
The question is: Will the result be better than an improved algorithm, or can the results from NavVis only be reached by using two scan heads.
My best guess is as follows, note this is based on personal observation of datasets from multiple sources and not rigorous academic research;

I think the word 'accuracy' can be a bit ambiguous here as there is much more to noise than accuracy. A multi-beam LIDAR on a SLAM system with the same distance accuracy as a single beam profiling LIDAR will exhibit far more noise. The reason for this is that the relative accuracy over very short distances on a profiler tend to be much better. This is because adjacent points on the same profile line have a very small time gap between them and hence will have very similar SLAM and poise parameters. Adjacent points on a multi-beam profiler are more likely to be interleaved, being observed from different beams, so while the mean accuracy may be the same, the standard deviation between adjacent points is higher and hence we see more noise (both in terms of position and intensity). From this, I'd speculate you potentially achieve some good results with the BLK2GO which you may not be able to with devices with multi-beam scan heads.

Once you use two scan heads you have extra redundancy and better coverage. Not sure what algorithm NavVis uses but there are a lot of ways to use this redundancy to improve data quality. What I have noticed in NavVis data is that the intensity in the processed scan appears to be synthesized from the photography. This suggests that the intensity is used for other purposes, possibly combined with colour to give relative weights to the same positional data from different LIDAR sensors when processing it. The processed data from the NavVis is made up of evenly spaced points, which I'm guessing are sampled from a smoothed model created from the combined LIDAR and photographic data. They've done a wonderful job here and it is for the other manufacturers to play catch up, which no doubt they will at some point.
User avatar
gsisman
V.I.P Member
V.I.P Member
Posts: 658
Joined: Fri Oct 07, 2016 1:51 pm
6
Full Name: Steve Long
Company Details: Montgomery County DOT _ MD
Company Position Title: Land Survey Supervisor
Country: United States
Skype Name: gsisman1
Linkedin Profile: Yes
Has thanked: 437 times
Been thanked: 97 times

Re: FARO ACQUIRES GEOSLAM

Post by gsisman »

smacl wrote: Tue Sep 20, 2022 5:04 pm
VXGrid wrote: Tue Sep 20, 2022 2:37 pmWhat I am woundering is, if the scan head is a lot better with less noise, if the resulting cloud is better as well.
I mean if the single point accuracy is better, the positioning part should be better automatically.
The question is: Will the result be better than an improved algorithm, or can the results from NavVis only be reached by using two scan heads.
My best guess is as follows, note this is based on personal observation of datasets from multiple sources and not rigorous academic research;

..... The processed data from the NavVis is made up of evenly spaced points, which I'm guessing are sampled from a smoothed model created from the combined LIDAR and photographic data. They've done a wonderful job here and it is for the other manufacturers to play catch up, which no doubt they will at some point.

I think Shane is right .There is some smoothing going on. I have several point clouds processed and delivered on pilot projects with the NavVis and the scan lines are very regular and spaced and all oriented the same direction ,though the operator did not stay oriented in a direction walking around everywhere, so the software is doing a very regular parsing /decimation and it makes it easy to view items from any angle in an orbit mode-much different from the BLK2GO which works great for seeing building and wall lines from a PLAN view.
If I remember multiple years ago from the Pegasus backpack days-also dual velodyne scan heads, one of the lidar heads was to help "AID" the IMU- like an early "VIS" technology just with Lidar points instead of photo pixels- I think that the VLX might work similarly, albeit with tighter specs and better cameras.

NAvvis data look.png
You do not have the required permissions to view the files attached to this post.
Post Reply

Return to “General Chat”