What accuracy can be achieved with MLS?
There are many advantages of using Mobile Laser Scanning (MLS) for corridor surveys including improved safety of operators, increased speed of data acquisition, comprehensive coverage of all features within the corridor and the high measurement accuracies that can be achieved. This article focuses on the ability of MLS to accomplish “Survey Accuracy”, for all points along a corridor, from MLS equipment operating at road speed.
When we developed our MLS system, we knew it was critical to meet the accuracy standards required by the Main Roads Department of Western Australia (MRWA). The MRWA specifications require a high standard of vertical accuracy for the hard surface measurements (i.e. on the bitumen or road pavement). The height of the road surface is critical when calculating the amount of material required to build the road. Even a small error in height (for example 3cm) on the road surface can have a significant impact on calculations for the volume of material required and the costs incurred.
The McMullen Nolan Group have developed MLS data acquisition and post-processing methodologies to achieve high accuracy road surveys. We have completed a range of road surveys across Australia that varied in length from 3km – 300km. Our results have been independently reviewed, and found to agree with field survey to better than +/- 15 mm. This high level of accuracy meets and surpasses the expectations of the Departments of Roads in the States where we have operated.
These high accuracies are achieved through:
- Calibration of the equipment before and after the survey
- The use of redundant measurements in data collection
- Creating a “Self-checking” control loop that compares existing controls, GNSS backbone and scanning, along the length of the survey
- Rigorous post-processing of the acquired data
Calibration of Equipment
The MLS calibration procedure takes place once the MLS equipment is assembled and before it is disassembled. The calibration procedure measures the relative position and orientation of the 3 major system components (GNSS, Inertial Measurement Unit – IMU and Laser Scanner). Without this precise calibration, point cloud data from a vehicle travelling along a road in one direction, will not be consistent with data collected from the vehicle travelling in the opposite direction. When accurately calibrated, the MLS will provide consistent results regardless of the direction of travel.
The calibration procedure also includes the measurement of GNNS antenna phase centres, if different GNSS receivers are used as base stations (see below).
Multiple Redundant Measurements
Incorporating redundant measurements and observations provides data strength and certainty to the survey. Redundant measurements include:
- Using additional GNSS Base stations along the route to measure the “backbone” of the survey and to provide multiple base stations for calculating the trajectory of the MLS along the route
- Measuring the corridor multiple times under different satellite windows to reduce the effects of satellite multipath, and changes in the satellite constellation
Redundant observations assist in identifying any sections of poor data (e.g. due to satellite obstructions), suspect control station values and any interference that may be in the point cloud (e.g. other vehicles). Comparing the output of redundant data sets provides a higher level of confidence in the final results.
The graph below displays the errors measured from six separate passes along a 15km section of road with an MLS scanner. Each colour represents a different pass. The error (predominantly GPS) for each pass in this test can be seen to vary from -0.04m to +0.03m. This section was recorded under ideal GPS conditions and during an ideal satellite window. Taking the “average” of all six passes reduces the position and height error contributable to GNSS to +/- 15mm.
Figure 1 – 6 passes of GNSS data over 15km
Self-Checking Control Loop
A self-checking control loop ties together all the control elements of the survey. A GNSS backbone is established along the survey route. This backbone is used by the MLS to determine the trajectory of the MLS system. The backbone is also used to check the existing control stations along the route. Targets are established on the road shoulder adjacent to the control station network. These targets are positioned with GNSS and leveled using standard spirit leveling to provide the correct orthometric height. While scanning takes place, the targets are encapsulated in the point cloud. The measurements of the control stations, targets and point cloud scan must be consistent with each other, providing a self-checking mechanism for the entire survey.
Figure 2: Self Checking Control Loop Methodology
Post-Processing Techniques
Rigorous post-processing techniques have been developed to ensure each stage of the survey provides consistent, high-accuracy data. Processing components include:
- Processing the GNSS backbone and target positions to provide high quality positions, with ellipsoidal and orthometric heights of all stations and targets
- Processing of the MLS trajectory, using both GNSS and IMU measurements for each pass, allows improved identification of any sections of compromised data
- Processing the collected data in both directions (forward and backward)
- Creation of a Point Cloud that incorporates all the collected data and is cleansed of any unnecessary data (passing cars etc.)
- Extraction of points and strings for the end-users’ CAD package
Figure 3- Extracted string data from MLS survey
McMullen Nolan Group are excited about the future of Mobile Laser Mapping and are pleased to offer a full service to end-users, or act as a “Mobile Laser Scanning” service provider to other Survey companies. Access to McMullen Nolan Group’s MLS will enable these Survey companies to offer their customers a greater range of technological measurement solutions to meet their needs.
For more information on McMullen Nolan Group’s Mobile Laser Scanning, please contact:
John Nolan (08) 6436 1559
Rod Eckels (02) 8084 5524
