We just completed a mobile LiDAR project that was designed to support a roadway resurfacing project in Orlando. The project was centered on the use of mobile LiDAR to generate roadway profile data that Engineers could use to design a resurfacing project. Obviously the data would need to be accurate and we were able to hit the mark and best of all – prove it!
Overview of SR417 Project
We collected the data using the new Riegl VMX-250 mobile LiDAR unit using a single pass in the north and southbound directions. We only required one pass in each direction to collect the road data which makes it very efficient from the data collection standpoint. In the past, we had to collect “strips” of data and then “sew” them all together during the calibration process. In this case, we took the opposing (NB and SB) strips and calibrated them relative to one another and then they are brought down to control as a final step.
LiDAR Coverage by Flight Line
Most of our clients are interested in the overall accuracies of the data, so we have built accuracy assessment tools that make it easy to review the LiDAR against survey control. The tool is simple to use and allows us to sort the results and dig deeper into the least accurate points to see why there might be discrepancies in the control vs the TIN surface.
417 Accuracy Control Report
For this project, we achieved an RMSE of .0525 ft – calculated by comparing the control elevations (Z) against the TIN elevations (Z TIN). This is important because we can check the point cloud against known control that was collected throughout the project and provide detailed information about the accuracy of the data.
Once the data has been calibrated sufficiently, we can then generate all of the derivative products for this project. We generated the following data for our roadway engineers:
- Pavement Cross-Slope
- Shoulder Cross-Slope
- 3D Roadway Markings
- Edge of Friction Course
3D Vector Data
This data set also supports detailed engineering analysis related to guardrail height above the roadway. This is an important factor to consider because there are specific standards that define where the guardrail is placed, more specifically, its height above the roadway, to corral vehicles that end up impacting the guardrail in an accident situation. The following graphic displays how this measurement can be made in the point cloud data.
Guardrail Height Measurements
Jason: This is a great article. Can you say what the survey control was set at. Is there a specific density or spacing that is suggested over flat terrain for producing Survey grade LiDAR. What was the survey control, RTK plus vertical leveling, using a base station?
Also if you had a VRS (Virtual Reference Station) to tie to the GPS to via cellular connection, would you be able to elliminate the need for survey Control?
We typically set survey control throughout the project on hard, flat surfaces that are easily identified in the LiDAR point cloud. Paint stripes and gores work best because their vicinity is pretty flat without many elevation changes in the area. This is important because we are mainly concerned with the Z elevations across the project to make sure that the overall surface met the project accuracy of 0.1 foot based on the control. Also, we had an independent surveyor run some cross-sections throughout the project to make sure the surface was within accuracy tolerances as well. We combined all of these “checks” into the final accuracy assessment report for the project and provided those results to the client.
Our survey control was set using Fast Static Setups (Trimble R8) and post-processed using the dual base stations set in the field. I also don’t think it is possible to eliminate survey control because we always need a way to “prove” to the client that we achieved a specific absolute accuracy across the project.
Thanks for the questions! Jason
Jason: thanks for the detailed response. What would a typical spacing between control and how far apart would you need cross sectional data. Can survey control be eliminated with VRS, can mobile GPS at highway speed use VRS?
Your Blog is the best on the web for LiDAR, please keep blogging about your innovative techniques.
I just sent you an example of a typical control layout that is the CalTrans MTLS standard that is being used for many of our mobile LiDAR projects. Just to clarify, the project we are discussing here only used the control to “validate” the absolute accuracy of the point cloud. The CalTrans method was used for another client interested in exploring another method of “transforming” the data to ground control – more of a surface fitting routine as opposed to letting the IMU carry you between GPS epochs. That being said, we aren’t using the VRS networks because our technique relies upon a static base station for the rover to post-process against using a very short baseline. The baselines for the VRS networks can be short or long, but this project was pretty small and we felt this technique could work best. Now, if you were collecting an entire County or State, VRS could be a much better solution – you just need to understand what errors you may be introducing into your error budget.
Ultimately, we try to think practically about the application of this technology. the CalTrans methodology provides a very high level of accuracy for small projects because of the heavy control that has to be set to Transform and Validate points. Base Stations work well when you can babysit the setup or put it in a secure area – you can set it up, collect and then break it down very efficiently. VRS is more cost-effective for much larger collection areas, so it is all about the application of the technology to fit the needs and budget of the project!