Merging into the Fast Lane
May 23, 2011
High-accuracy helicopter and terrestrial mobile LiDAR datasets have proven to be accurate sources for engineering-grade digital elevation models. However, LiDAR data acquisition is a complex, highly innovative and rapidly evolving field. Because of the varying levels of accuracy and resolution that can be obtained, relying on one type of data acquisition is not usually sufficient for large-scale transportation or infrastructure projects. The combination of two or more methods is increasingly recognized as a best practice among service professionals, but merging the datasets from multiple types of methods has been challenging due to software limitations and other hurdles. Additionally, there has been a broad lack of understanding on the part of government agencies as to what results can reasonably be expected through the use of these advanced technologies.
In late 2010, the Texas Department of Transportation (TxDOT) embarked on a study of a 2,500-foot section of Interstate Highway 30 (IH 30) in Mesquite that was experiencing repeated water accumulation. The project presented TxDOT with a unique opportunity to closely evaluate the use of LiDAR as a suitable technology for data acquisition. “We wanted to figure out how much water was accumulating and how it was accumulating on the surface,” says Mark Eder, district survey coordinator for TxDOT. “We needed some very precise data, so we decided LiDAR was the answer.”
Eder already had some experience with terrestrial LiDAR and had been studying mobile mapping, but he wanted to take this project a step further and incorporate mobile mapping with helicopter LiDAR. “By evaluating the product produced by terrestrial, mobile and helicopter scanners, we hoped to gain a better understanding of the accuracy of each type of scanner so we could begin developing standards for the TxDOT survey manual,” Eder explains.
The agency decided that teaming several expert firms in the LiDAR field would provide the best approach. Woolpert, Aerial Data Service Inc. (ADS), Surveying and Mapping Inc. (SAM) and Tuck Mapping Solutions were contracted to collect and process datasets for the project using two Optech Lynx Mobile Mapper systems and a Riegl LMS-Q560 full waveform 164 KHz airborne laser scanner. “We had never deployed all of that equipment on one project before, much less tried to merge all the data together,” Eder says. “We didn’t really know what to expect.”
Advances in technology have brought about significant changes in large-scale mapping projects. Traditional methods of highway mapping have required a large number of “boots on the ground.” Analog cameras have produced imagery for mapping for more than 50 years and have achieved very good accuracies when placed in low-flying helicopters. However, a large number of targets must be placed along the active roadway and surveyed with digital levels to provide high accuracy control--a requirement that is both time consuming and expensive. Projects involving bridges, areas of dense tree cover and the mapping of curbs, gutters and manmade features have required detailed surveys that could only be obtained on the ground. This restriction presents safety concerns for the survey crews working in and around traffic and often requires costly lane closures.
The introduction of LiDAR units to low-flying helicopters and, more recently, moving vehicles, has allowed mappers to cover many miles of roadway each day while also reducing costs compared to conventional surveys of the same area. In the past, surveying and mapping a 30-mile segment of road might have taken months. Today, this data can be acquired in a few hours and processed in a matter of weeks. Maps are also being produced with record accuracies; it is not uncommon to map a large segment of highway with an accuracy of .05 to .1 foot. This accuracy is allowing design engineers to provide construction volumes that have fewer challenges to quantities during construction.
Mobile and aerial LiDAR methods are very complementary. While mobile mapping produces a survey of features from a horizontal perspective and is ideal for capturing vertical objects in the right-of-way (especially retaining walls and barrier walls), airborne systems provide the ability to map a wide corridor outside of the right-of-way, including subdivisions, malls, businesses or other building structures that will be affected directly by the construction or may be affected indirectly by noise, dust or other effects. Most corridor maps are 500 to 800 feet wide, but it is not uncommon for engineers to request maps of areas that are up to 1,000 feet left and right of the alignment, which can be easily achieved with aerial systems. By using an aerial system to capture high-resolution color imagery at the same time as the LiDAR point cloud (with both sensors matched for width), digital elevation models (DEMs) can be generated for the creation of the digital ortho for the project. The resulting maps include an adequate density of LiDAR points on the ground under trees to depict the bare earth--even when the mapping takes place with full foliage in very dense tree areas. Aerial LiDAR can also capture power lines, including the sag of the catenaries, and it can provide the level of detail required for planimetric mapping.
Conversely, mobile mapping provides details of drainage structures such as curbs and gutters as well as breaks in the pavement--details that are difficult to capture from the air. When used with software such as TopoDOT from Certainty 3D, mobile mapping allows cross-sections anywhere along a corridor of dense point cloud data to be easily extracted for use at the engineering design and operations level.
Mobile mapping also provides for images to be registered to the LiDAR points, thus defining the roadway features in a way that was not possible in the past. By colorizing the LiDAR point cloud, an exact fit between the LiDAR point cloud and the imagery can be achieved. (Typically it is preferable to map point cloud data to the same view as the image overlays, thereby using full camera resolution for feature identification, extraction and modeling operations.)
One of the most remarkable features of mobile mapping technology is the ability to produce a high point density on objects in the roadway. Some mobile LiDAR systems can produce up to 10,000 points per meter, with 5,000 points per meter being a good average. This high density of points provides detail that has never been achievable with conventional ground surveys. If the designers have a question about a critical area of design, they can rely on the point cloud to provide the required information.
Mobile mapping systems require ground control to be surveyed within the right-of-way and within the sight range of the mapper as it travels down the road, but this is one area that is expected to be refined in the near future. Improved methods of ground control will reduce the amount of surveying needed in and around the roadway.
Using a combination of mobile mapping and helicopter LiDAR, the project team working with TxDOT collected all of the data required from the 2,500-foot stretch of IH 30 in less than one day. No lane closures were required, and except for placing the control (handled by TxDOT Dallas District), no personnel were on the roadway, which provided a substantial safety benefit. The next step was to process and merge the mobile and aerial datasets to produce the final deliverables.
In the past, merging multiple datasets has presented significant challenges due to variations in the output file format and accuracy levels. However, vast improvements in both technology and software have occurred within the last several years. Today, most mobile mapping systems and airborne systems use LAS files as the standard for data output. As a result, data from multiple systems can easily be merged into a single point cloud. Since mobile mapping uses GPS and the inertial measurement units (IMUs) that are common with airborne systems, accuracies are similar. In the TxDOT project, the mobile mapper accuracy ranged from .04 to .08, and the helicopter mapping accuracy ranged from .04 to .10, which was comparable to the .03 to .05 accuracy obtained from terrestrial scanners. “We were really surprised by the accuracies of the mobile and helicopter systems,” Eder says.
Combining the point clouds obtained along IH 30 improved the quality of the data and allowed the synergy between the aerial and mobile datasets to be exploited. Within TxDOT, Eder used Autodesk CAiCE Visual Transportation software to merge the datasets and create a seamless digital terrain model (DTM). Because this was an evaluation project, the consultants also processed the data and created deliverables. For Tuck Mapping, Riegl USA and Certainty 3D assisted with merging the datasets and constructing a DTM.
The mobile mapping data provided a high level of detail where it was needed, such as on the pavement, where several thousand points per meter were collected. With the addition of the airborne mapping data, a cross-section was extended beyond the right-of-way. Such detail is especially important when new construction is being planned and when roadway slopes will be modified during construction.
Retaining walls, bridge abutments and other tall structures could be seen from all angles when the datasets were combined. Likewise, bridge beams could be mapped more accurately when the mobile mapping data acquired from the superstructure of a bridge were merged with a profile of the top of the bridge acquired by helicopter LiDAR.
The TxDOT project provided ample evidence of the utility and synergy between airborne and terrestrial mobile LiDAR data as well as its superiority over traditional survey methods. Enough points per area were acquired to extract all features of interest, and the information was easily identified, extracted and modeled. What’s more, the data can now be easily accessed and used across engineering design and operations.
Based on the results of this project, it is clear that combining mobile and aerial LiDAR technologies can provide accelerated schedules, increased safety, streamlined operations and improved quality for highway design projects. “This project has definitely had an impact here in Texas,” Eder says. “Now that we’ve had a taste of the technology, it’s going to be utilized more and more. It gets people out of the roadway surface, so it’s safer, and the datasets are very accurate. We’re already using mobile mapping for some projects, and the combination of mobile and aerial will come into play on future projects. The data that we didn’t have before is now easy to gather with this technology.”