Surveying Road Surfaces with LiDAR
In this post we are going to take a look at how Rekon uses their LiDAR system to make roadway surveying safer, faster, accurate and cost effective for our clients. LiDAR (light detection and ranging) has been around for decades but, in recent years the size and cost of the equipment have steadily been dropping whiles the range, accuracy and pulse rates have increased pound for pound.
At Rekon we deploy a LiDAR system made by Routescene. Benefits of the system include a stand alone design, light weight construction, ability to be mounted on a drone, vehicle or even walked, and a great customer support team. With this Unit we are able to produce 700,000 pulses per second. With the integrated INS (inertial navigation system) we are able to take those points and position them in road scans to a degree of accuracy equal to that of a terrestrial GPS survey.
The Lidarpod Vehicle Mounting system used for this operation was supplied and designed by Routescene, the manufacturers of the LidarPod. It's main function is to equip a vehicle like the ½ ton pickup in the picture with the mounting points for the Lidarpod, the GPS antennas, and the wheel mounted odometer. The main truss system is isolated from vibrations with a shock absorption system.
In this case the mapping area was defined as 3.5 kilometers on Okanagan Centre Road E from Oceola Road to Berry Road. The client required the current road surface with points every 5 meters along the road alignment at the center and edges of pavement. A vertical accuracy of 2 to 3 cm was required.
As with terrestrial GPS surveys, a base station communicating correcting in real time or for post processing is required. In this case, the project was inside of the Can-Net subscription network so, setting up a base station like the one seen in the picture was not needed. In areas where access to corrections from a CORS (continuously operating reference station) is not available a local base can be deployed. This local system is designed to directly communicate over a 400 mHz band width with the LidarPod to provide real time kinematic correction or to be used in post, where log files are downloaded and processed to correct telemetry data from the LidarPod.
To ensure the LiDAR data is aligned in the X,Y & Z as well calibrated in the roll, pitch and yaw, laser targets are placed in the scan area strategically. These targets, are leveled and surveyed using terrestrial survey methods and they serve as monuments in the LiDAR scan that
can be evaluated for position, elevation and orientation. The targets are easy to locate in the scan because each point in the LiDAR scan reports intensity along with position, since the targets are highly reflective, they show up like signal flares.
During the scanning process the configuration of the Lidarpod, the status of the INS, GNSS and system vitals are monitored to ensure that all is operating as it should. The LidarPod outputs this constant feed of status information via a USB cable that is routed into the cab of the vehicle and displayed on a computer.
The only time that a person had to get out and work on the road grade for this survey was to set out the targets and to record terrestrial GPS data as part of our QA process. That meas that our method reduced the human to traffic interface substantially even for a short project like this one. Target locations and QA points can also be placed and planned for in areas of little to no traffic, further reducing the hazard of working near traffic.
The best times to scan a road surface is during low traffic volumes and never during rain, snow or fog. Scanning can happen during the night when traffic is low since it does not require day light. During this scan a few cars were captured in the scan but, were easily removed by profile trimming them out manually during the processing of the point cloud.
After the road scan, the navigation trajectory data was post-processed using Kinematica. Kinematica is used to take the trajectory data from the Lidarpod and post-process the location and orientation. This gives a more accurate location and orientation for every second in the LiDAR data. The trajectory data in a LiDAR scan defines the locations of the laser scanner in 3-dimensional space. Increasing the accuracy of the trajectory data (commonly called track data) directly increases the accuracy of each LiDAR point.
LiDAR raw scan data and trajectory data that was produced by post processing in Kinematica were then opened using the LidarViewer software. In this software we combined the points with the trajectory and then narrow the selection of points that are to be exported based on range, intensity, position along the trajectory, and field of view. How and what points are filtered was and is always based on project requirements, things like accuracy, target area, deliverables and required point densities will contributed to deciding the parameters during this process. Once we were done here a 3D point cloud of the scanned area is exported in native LiDAR point cloud formats like .las or .laz.
According to the configuration parameters applied to the LiDAR and the planned driving speed, the theoretical points density was planned to be 1487 pts/m². This counted on a speed of 50 km/hr. in one pass of the mapping area with the scanner speed at 1200 rpm (20hz). The average actual LiDAR coverage density during the scan was 1476 pts/m² on the road surface. The total area covered during the scan was 27.6 Hectares producing a total of over 40 million points.
As part of our QA process the elevations measured by the LiDAR point cloud and the terrestrial GPS survey points were compared to find the difference. The error distribution, shown in histogram resulted in a RMSE (root mean square error) of 0.027m with a standard deviation of 0.026m. With these results we were within our clients requirements for accuracy.
Finally using Global Mapper and AutoCAD Civil 3D we produced the deliverables for this project. The deliverables included; a DEM (digital elevation model), a point cloud, and an AutoCAD surface of the roadway, all on a one by one meter grid. Additional deliverables could have also included, deliverables such as; road lines in a shapefile or other vector file formats, clearance reports on overhead utilities, points files locating features in the road like manhole covers, and even locations of defects in the road surface such as cracks and potholes.
This project was a relatively short run compared to the amount of ground that can be covered per day using the vehicle mounted LiDAR system. For more information on what Truck LiDAR can do for you and your project Contact one of the experts at Rekon.