Laser Scanning – Best Practices
Does anyone know where I can find a document of some kind for field personnel, and also for project managers, defining Best Practices for laser scanning.
I’m looking for something saying how many targets should be in overlapping scans, how many control points for georeferencing, how far to keep the truck away from the scans, that sort of thing.
We’re just starting out with a Trimble X7 scanner and I’m having way too much trouble getting my scans to register to each other.
We do mostly highway work, and I’m having trouble finding enough well defined features for registration.
I’m also looking for the same type of thing for drone work, with the Rock Robotics R2A.
- 43 Replies
question: how far between scanning stations are you advancing?
one thing you could do would be to use a shot point on the ground with a movable object like a hollow oblisk that you could get some good orthogonally oriented returns between structures or signs. more work yes, but consistent and a known dimension you could calibrate to.
never used the x7 but I have used the SX10-12 and using my traverse points never ran Into registration issues that couldn’t be resolved unless there was a set up bust and that just took some additional work to get it align.
the whole automatically setup thing always makes me feel like it’s too ‘black box magic’ and has issues with the actual deployment and success in mission planning.
my$0.02
Highway work and x7 in a same sentence = recipe for disaster.
This scanner is a small scanner to do mainly buildings were lot’s of reference planes are visible, not so with highway jobs.
C2C = impossible so targets are needed.
Imho they sold you the wrong scanner.
question: how far between scanning stations are you advancing?
one thing you could do would be to use a shot point on the ground with a movable object like a hollow oblisk that you could get some good orthogonally oriented returns between structures or signs. more work yes, but consistent and a known dimension you could calibrate to.
never used the x7 but I have used the SX10-12 and using my traverse points never ran Into registration issues that couldn’t be resolved unless there was a set up bust and that just took some additional work to get it align.
the whole automatically setup thing always makes me feel like it’s too ‘black box magic’ and has issues with the actual deployment and success in mission planning.
my$0.02
Somewhere between 40 and 60 feet.
This is only the third job I’ve been working on, I’ve already told the higher ups that we need more target points, I’m just looking for something in writing to back that up and quantify it.
I was actually all for the SX12, for the types of jobs we’ve been doing with it so far I thought it would be fine. Plus the fact that you can actually have control on your setup points. And I’m also not crazy about the black box effect. It seems to make everyone think you barely need control which is of course not the case.
Highway work and x7 in a same sentence = recipe for disaster.
This scanner is a small scanner to do mainly buildings were lot’s of reference planes are visible, not so with highway jobs.
C2C = impossible so targets are needed.
Imho they sold you the wrong scanner.
Well we’re mostly just using it for pavement tie in type jobs at this point. Couple hundred feet or so.
I expect people are going to want to go up from there but we’re really going to need a lot better procedures in the field for this to start making us any money.
The wrong scanner? Maybe. We definitely didn’t get enough training. Especially for the field guys.
using the sx10-12 we found a sweet spot between 150 to 200 feet for coarse scans to get good time and productivity. we have a water tower that after 10 or so scans from various traverse points started to appear in the default layer and it was more than a 1/4 miles away.
Picked up birds, planes, and the coolest ones were the ascending and descending helicopters.
they are great tools when you get a good workflow.
there’s probably a use case and workflow that can bring it together for your team. make the sales guys get out there and earn that commission!!!
Highway work and x7 in a same sentence = recipe for disaster.
This scanner is a small scanner to do mainly buildings were lot’s of reference planes are visible, not so with highway jobs.
C2C = impossible so targets are needed.
Imho they sold you the wrong scanner.
I agree with this. I have used the X7 and it is a great scanner but with Perspective alone it’s not for highway work. The cloud to cloud registration is more suited to buildings. You can use the X7 for what you are doing but you will have to use targets and get a copy of Realworks.
Moving such short distances for highway/corridor mapping sounds agonizing. I like the X7 a lot, but for some weird reason it seems to be applied where it shouldn’t. We struggled with this too, mostly because the powers-that-be didn’t really pay attention when purchasing newer gear.
At the very least I would want an X12 (big bro to the X7) or equivalent, and if you’re doing lots of this type of work as you suggest, I’d seriously be looking at mobile LiDAR.
It looks like you have sUAS, in which case maybe an SX10/12 is a better fit because most of what you’d be doing is infill after collecting the UAS data.
“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil PostmanYou can also mount the Rock 2A on a vehicle and drive the route
You can also mount the Rock 2A on a vehicle and drive the route
I’ve been suggesting that. Haven’t gotten around to it yet.
You can also mount the Rock 2A on a vehicle and drive the route
Do you have a R2A? Maybe you could answer a question I’ve had.
I really haven’t had time to look into documentation on this, since I haven’t been doing any of the processing on the drone data, but I was wondering if we got a mount for a truck to use with the R2A, would the boresight procedures need to be redone every time we switch from the M300 to the truck mount and back?
I have a R2A, but have not tried it on the vehicle yet. My understanding is that it is interchangeable between the UAV and the vehicle. You still need to do a 5 second acceleration and figure 8 to calibrate it.
The problem with scanners that use targets inside the scans is, each scan is aligned to the next, errors accumulate over many scans, if you try completing a closed loop, without overlap between scans on the opposite side of the loop, you’ll see what I mean. What you need is a long range scanner with known occupation positions and you’ll also want a long distance backsight target, otherwise there will be rotational alignment errors.
You are probably best off using a scanning total station to capture a low resolution cloud, then you can use point cloud alignment software (Vercator can handle natural environments) to place and align your X7 scans into the point cloud captured by the total station.
Cheers,
Peter.
@bobwesterman The SX12 is the right tool for the job, if you want high resolution point clouds, then you can use the X7 with the SX12, but you can’t use the X7 without the SX12. Try taking the X7 back and swapping it for an SX12. The SX12 will only give you low resolution point clouds, but that’s probably all you really need anyway, better to be accurate with less data, than a huge amount of data with no accuracy (if the X7 point cloud is not accurately geolocated, then none of the point coordinates will be accurate).
The problem with scanners that use targets inside the scans is, each scan is aligned to the next, errors accumulate over many scans, if you try completing a closed loop, without overlap between scans on the opposite side of the loop, you’ll see what I mean. What you need is a long range scanner with known occupation positions and you’ll also want a long distance backsight target, otherwise there will be rotational alignment errors.
Will control on to some targets throughout the survey not take care of this?
@totalsurv It depends on the accuracy you’re looking for. Lets say for example you have control points every 10 scans, you’ll still require at least 3 targets common between each scan, there will be some variance between the control points and the scanned road surface my deviate up or down between the control points and the length between control points and the combined length of the aligned scans may differ, as a combined point cloud is aligned to the control points after scans are aligned to each other. Also if your doing lower resolution scans, the targets will need to be relatively close to the scanner, which limits the spacing between scans and affects angular alignment. Wrong tool for the job, it’s better suited to buildings, man made structures eg industrial plant, where information in scans means you don’t require many targets, but the loop closure problem still remains and one day will catch you out when you least expect it, if you’re not aware of it.
@bobwesterman The SX12 is the right tool for the job, if you want high resolution point clouds, then you can use the X7 with the SX12, but you can’t use the X7 without the SX12. Try taking the X7 back and swapping it for an SX12. The SX12 will only give you low resolution point clouds, but that’s probably all you really need anyway, better to be accurate with less data, than a huge amount of data with no accuracy (if the X7 point cloud is not accurately geolocated, then none of the point coordinates will be accurate).
If the X7 is struggling to register two of it’s scans cloud to cloud then how is it going to be better to try to register it to a lower resolution from a SX12? Is Vercator that good that it can cloud to cloud scans that Perspective (which is the best cloud to cloud I have used) can’t cope with even with little alignment planes and vegetation etc.?
@totalsurv It depends on the accuracy you’re looking for. Lets say for example you have control points every 10 scans, you’ll still require at least 3 targets common between each scan, there will be some variance between the control points and the scanned road surface my deviate up or down between the control points and the length between control points and the combined length of the aligned scans may differ, as a combined point cloud is aligned to the control points after scans are aligned to each other. Also if your doing lower resolution scans, the targets will need to be relatively close to the scanner, which limits the spacing between scans and affects angular alignment. Wrong tool for the job, it’s better suited to buildings, man made structures eg industrial plant, where information in scans means you don’t require many targets, but the loop closure problem still remains and one day will catch you out when you least expect it, if you’re not aware of it.
I never said it was the right tool for the job. It is however possible to use targets and control to get a tight georeferenced and verifiable registration with the right procedures without a scanning total station. Otherwise our scanners are useless.
@totalsurv Yep, if you like measuring lots of targets and keeping track of them all, you can make it tighter, but still not as accurate. Not really worth the time though. The X7 and other scanners like it don’t really like big open spaces.
While you can get spherical targets with prisms, I would make a small window scan of each target using a scanning total station, then you’ve got something for the X7 to register to, that’s if you want high res scans, if not, I would use the scanning total station.
Yes, those scanners are useless in some situations, but useful as others have described, a hammer isn’t a very good spanner, but you can get the nut loose using a hammer and a chisel,with a little heat from an oxy, but you would only do it that way if you had to. I think this is the same situation.
The problem is, when you leave the job site, you have a lot of scans that are yet to be aligned, and if it doesn’t work out, well, you’ll have to return to the job site and set up the targets in the same places etc.
With the scanning total station, you already have a low resolution fully aligned point cloud when you leave site, so you are much less likely to have to return to the job site.
I hope that puts it into better perspective.
Cheers,
Peter.
The targets issues are a reason why i bought a Riegl Vz400i 5 years ago to stop fiddling about with spheres and targets with my old Trimble TX5.
GNSS on board and the best c2c algorithm never lets me down. But is is a lot more expensive then a x7.
@beuckie Interesting scanner, I haven’t had the opportunity to use one. Guessing you’d be using RTK with a base station, so c2c only has to rotate the clouds into alignment.
Log in to reply.