The Transportation Research Board Committee for Geospatial Data Acquisition Technologies in Design and Construction recently held its summer meeting in Philadelphia in late July. A portion of the meeting included presentations and discussions concerning mobile mapping. Typical of some forums, the discussion generated opinions and opportunities about this emerging technology. I personally find the discussion more genuine when both the advantages and disadvantages of a technology are discussed.
Stating that mobile mapping is an emerging technology might be an understatement. Lewis Graham with GeoCue reported there are roughly 265 mobile mapping units in the US, but only 15 would be considered metric quality units. Metric quality would be those units being used to support the engineering applications for design. The remaining units would be those more closely aligned with GIS and asset management.
Much of the discussion concerning mobile mapping centered on accuracy, precision and the associated costs. The difficulty with this technology is attempting to make sure end users understand the limitations in accuracy and precision involved with mobile mapping. Various techniques and technologies can lead the end user to erroneous conclusions. The massive amounts of data collected by the non-metric quality units typically produce the high precision but lower accuracy results. It is easy to view the data with the high precision and think high accuracy goes with it based on the massive amount of points collected. Lewis also expressed an opinion that lot of folks may be disappointed when they try to leverage GIS quality 3D data for high accuracy work.
One of the interesting bulleted points in a presentation by Lewis was that the lower accuracy data from these units may provide 75% of the geospatial information that users actually need. Related issues of data volumes and accuracy vs. precision were also identified. Such topics spurned lively discussion as would be expected. Lewis pointed out that in his opinion, mobile LiDAR data collection for highways should be focused on collecting high desnsity data that, with the introduction of supplemental control, could support design applications. Mobilization, processing and supporting end user applications are the costly components, not the collection time or data storage requirements.
These comments made me reflect back on many other geospatial technologies that I have been involved with during my career and the associated learning curves for each. I believe the debates will evolve over the mobile mapping technology the same way they did with prior geospatial technologies. Over time the educational aspects will result in the users becoming more aware of the related accuracies and tolerances. But this wont occur without some growing pains. As with many technologies, enhancements to the hardware and software will continue to improve collection and end user skillsets.
When digital aerial cameras started out, the issues about engineering quality and associated applications were hotly debated. We had the push broom sensor versus the digital frame type sensors. Along with this debate was whether digital imagery could match the quality of the film based cameras. Some questioned if photogrammetry was about to be replaced with Lidar. Many of those companies involved in the original aerial sensor race have fallen by the wayside and digital cameras providing the higher quality outputs have seemly become king of the hill. Both airborne Lidar and photogrammetry are now tools of the trade.
With GPS we again had the misunderstanding about precision, accuracy, and cost between handheld, resource grade and geodetic receivers. Many learned there was no easy button and massive amounts of data failed to provide the information necessary for engineering design. I expect a whole lot of data was collected, and only a small portion of it became usable for engineering design applications. But without a doubt, persuasive cases will be made that due to the density of the data and its associated precision that we can live with less accuracy.
There will also be persuasive arguments that bad data is better than no data. As with GPS, it is important to know how good or bad the data really is. This brings us back to the issue of accuracy vs. precision and perhaps to the need for a requirement that a statement accompany the data concerning its lineage and intended use. This metadata should include a specification of the accuracy and precision.
I believe as we work our way through the mobile mapping learning curves, this tool will be added to the arsenal of tools for the engineering map makers, but this time I hope we are smarter about managing data quality.
As an engineering quality map maker, I find the need to deliver faster and better data unsettling. Little errors in measurements and position are so difficult to detect. Vertical differences of a few tenths of a foot blend so nicely into the mapping and the digital terrain models. Larger measurement differences are easily spotted and some corrective action taken. Best surveying practices will need to become standards for the data to be equal in quality. When users of the data understand the error budget with regards to the data, then we can decide if it is actually better data.
In transportation design, I often hear comments about the survey data being close enough for road building, and we dont need the same accuracy as if we were building a piano. A couple of simple questions need to be asked of those with that opinion. What are the required standards for design and construction? Both design and construction have differing needs. However, if you are the contractor bidding and building a roadway project and want to use automated machine guidance (AMG), your requirements are going to be for great precision and accuracy for roadway projects.
Those contractors wanting to use automated machine guidance will have a completely different set of needs than the designer. The requirements for greater precision and accuracy are paramount when contractors are using AMG for both grading and surfacing. The quality of the data will be directly reflected in the estimated quantities being bid by the contractor. To gain the full benefits of automated machine control, contractors are keenly aware that time to complete the work, pavement smoothness incentives, and the estimated building material quantities for the project are major items of a contract.
With all the survey requirements necessary to meet the design requirements, and the movement toward three dimensional design mobile mapping will clearly emerge as a desired technology tool. But to me, it seems like the mobile mapping folks are not as acutely aware of the issues regarding design quality data. I suspect this is from a lack of understanding about the requirements for design. I believe this is supported by knowing there are so few metric quality units being used. The best solution always seems to circle back to an earlier point about collect it once, get high quality data, and use it as you need. With quality data, the users ability to use the data repeatedly for other applications may result.
Another benefit provided with metric mobile mapping units is having color imagery, color associated with lidar point data, and quality measurement data as deliverables. All of these products are key elements associated with advantages of 3D design. Point clouds in a two dimensional CADD environment or a three dimensional stereo system enhance the designers visualization of the design. To further illustrate understanding design considerations I point to a portion of American Disabilities Act (ADA). Color of signals, placement and color of pedestrian buttons, slope and dimensions at crossings and in front of pedestrian buttons, and crosswalk slopes and roadway grades are features needing to be surveyed.
It may take some time to work out all of the details and standard operating procedures but LiDAR point clouds from either a static or mobile platform appear to be the tools of choice for many transportation survey applications.