- How much is accurate lidar surveying data for engineering design projects like road, railway, dam and irrigation??ÿ
- How much is accurate lidar surveying data for engineering design projects like road, railway, dam and irrigation?
Short answer, +/- 5-10cm
Long answer, it depends...
Are you asking how much it costs?
I would argue that you can get within 3cm or perhaps even less, but depends if you are talking about terrestrial lidar, mobile lidar, drone-based lidar, etc. Need more project specifics....
I would argue that you can get within 3cm or perhaps even less, but depends if you are talking about terrestrial lidar, mobile lidar, drone-based lidar, etc. Need more project specifics....
sorry I read the op as aerial LiDAR for some reason. Yea terrestrial can definitely get below 3cm. But again, too many variables to answer this question. It’s like asking if a car can go 250 mph.
sure, some can. Your old beat up civic? Maybe not.
We recently spent about $45k on a DJI drone and associated softwares. We have tried to use both Photogrammetry and LiDAR to create a deliverable drawing. Grant it we are still in the rookie phase, if not still training camp, but I am a bit disappointed in the results and workflow and software capabilities.
There seems to be about a dozen major softwares that you can choose from and just about whatever you choose you will need 3-4 of them to get the job done.
The need to jump from one software to the next is a quite annoying. We use DJI Terra to get the raw data processed. Then Pix4d survey or Pix4d Matic to further process. Then if you want a point cloud in civil 3D you have to first run it through autodesk ReCap. A lot of steps and variables.
The software claims of feature extraction (for surface breaklines or any other desired linework) is very overpromised from my experience so far. Drawing them in the drone softwares is a major pain. I think it will be another 5 years before this feature works consistently, reliably and accurately.
In the field thus far we have also shot many if not all features with total station or GPS as a check for the drone data vs more conventional methods. I’m seeing horizontal and vertical discrepancies of .25’.
And at the end of the day I have an uneasy feeling about the data. But like I said, we are still rookies. I can pretty comfortably say I do not think the technology as a whole is as far along they claim or we were hoping.
I am not giving up on the whole concept of using drones. At this point in time, I think they need to be used carefully on a project by project basis. Integrated with a lot or a little of GPS or Total Station shots.
For a vast majority of the projects we do I am not seeing drones as a time saver. All that passive processing time. All that active processing time. Drawing breaklines in a unfamiliar drafting platform. Integrating both drone data and conventional data is always a pain. I think at this point a boots on the ground survey delivers you much more confident results in nearly the same time.
We have one long standing contract for surveying volumes of stockpiled sand. We have not used the drone at that site yet but I think this is where the drone will shine. And save me from walking up and down 100’ piles of sand.
We recently spent about $45k on a DJI drone and associated softwares. We have tried to use both Photogrammetry and LiDAR to create a deliverable drawing. Grant it we are still in the rookie phase, if not still training camp, but I am a bit disappointed in the results and workflow and software capabilities.
There seems to be about a dozen major softwares that you can choose from and just about whatever you choose you will need 3-4 of them to get the job done.
The need to jump from one software to the next is a quite annoying. We use DJI Terra to get the raw data processed. Then Pix4d survey or Pix4d Matic to further process. Then if you want a point cloud in civil 3D you have to first run it through autodesk ReCap. A lot of steps and variables.
The software claims of feature extraction (for surface breaklines or any other desired linework) is very overpromised from my experience so far. Drawing them in the drone softwares is a major pain. I think it will be another 5 years before this feature works consistently, reliably and accurately.
In the field thus far we have also shot many if not all features with total station or GPS as a check for the drone data vs more conventional methods. I’m seeing horizontal and vertical discrepancies of .25’.
And at the end of the day I have an uneasy feeling about the data. But like I said, we are still rookies. I can pretty comfortably say I do not think the technology as a whole is as far along they claim or we were hoping.
I am not giving up on the whole concept of using drones. At this point in time, I think they need to be used carefully on a project by project basis. Integrated with a lot or a little of GPS or Total Station shots.
For a vast majority of the projects we do I am not seeing drones as a time saver. All that passive processing time. All that active processing time. Drawing breaklines in a unfamiliar drafting platform. Integrating both drone data and conventional data is always a pain. I think at this point a boots on the ground survey delivers you much more confident results in nearly the same time.
We have one long standing contract for surveying volumes of stockpiled sand. We have not used the drone at that site yet but I think this is where the drone will shine. And save me from walking up and down 100’ piles of sand.
In my opinion if you want to do lidar with a drone or aerial there is only one solution : Riegl. All the rest fades away compared to them. But pricing starts at a minimum of 150k.
DJI can makes drones but there it stops. Lidar is a different ballgame.
@beuckie what is it with Riegl that makes them superior in your opinion?
@beuckie what is it with Riegl that makes them superior in your opinion?
Family owned company, no Hexagon or other multinationel behind them
Excellent r&d
Produce lidars for various applications
Zenithal-Nadir scanners
They only do lidar, nothing else
...
@ryancj31 Thanks for the info. I think it is the same with terrestrial lidar. Many people say you should scan such and such a project but when it comes down to it processing the data and extracting meaningful and accurate information is incredibly time consuming. Like you say most of the time you would be just faster and more accurate with traditional survey gear. I don't doubt that you will find use for the lidar you have purchased but it think you are realizing it is limited to certain projects and may take longer than expected to get return on investment. Keep us updated on how it goes over next few months.
I have had good luck with Virtual Surveyor for extracting elevation data to be imported into CAD. You start by importing a point cloud or DEM (and ortho, if you'd like). After that it is very easy to draw breaklines and create grids of points. You can even remove unwanted features from a point cloud with just a few clicks. Exporting is easy too. I usually export the breaklines as a dwg and the points as a CSV. Then I import them into CAD and create a surface as I normally would.
Gregg
@beuckie Thank you for your experience based reply. Here i have attached the discrepancy between lidar and ground surveying data. Very annoying. Can you see it and tell me the probable reason? Thanks
There's no way to "diagnose" the problem simply by looking at a spreadsheet of the deltas.
Those are some large vertical discrepancies, even on the GCPs, which is extremely suspicious. Our LiDAR work turns up a few cm, maybe a decimeter at the very worst, and it's generally always a bias in one direction so the cloud is simply bumped up or down to align it.
It's also super suspicious that every single one of the GCPs has zero difference horizontally. There's always going to be some difference unless some seriously weird rubbersheeting is being applied.
Why not ask the LiDAR provider to either fix their product or offer a detailed explanation of why the vertical is so far out versus ground-truthing? Surely there is a spec (ASPRS or NSSDA) that they are required to meet per the contract
@rover83 Alreday I have asked the provide. But I am not satisfied with their response. I bring it here just for discussion and share experience with other professionals. Do you want to see the `*.las file?
There's a lot of elements and factors with why the point cloud wouldn't match the conventional data. Not easy to determine from what you're showing us. I would want to know about what equipment was used, methods, mission details, etc.
Are you comparing to a control network that has been established with a level for your vertical component or are you perhaps comparing to GPS/RTK?
If the provider's answers to your concerns aren't progressing towards a solution or simple conclusion, that's not great. Significant QA/QC should happen before data is delivered. The equipment might not be good enough for the tolerances on hard surfaces,(I.e. centerline road, structures, etc.) like the guys mentioned this above because the less expensive scanners have a greater vertical range in the point cloud. This data needs to be cleaned up considerably compared to resulting point cloud from the Reigl sensors.
To me, it looks a lot like somebody exported the point cloud without any cleanup or removal of noise and outliers. Exported points from a DSM perhaps, not cleaned up bare-earth. There seems to be some okay data in there, perhaps with a Z-bias that could easily be corrected but it's impossible to say without information on what they're doing to arrive at the .csv they sent you.
If you wanted to zip it up and send me some info about the coordinate system, I could take a quick look. I realize it's been two weeks since your post.
