| Operational Visualization Applications for Decision-making at the Texas Department of Transportation Using LIDAR Data
By John M. Hill, Cindy Gloyna, Kathleen Chavez, and Dana Young Introduction
Houston is the largest city in Texas and the fourth largest city in the United States. The population of the Houston metropolitan area is expected to double by 2020. According to a recent study, the Katy Freeway corridor - a segment of Interstate Highway 10 (I-10) - is one the most congested arterial highways in the region. The American Highways User Alliance identifies the I-10 at I-610 loop as one of the ten worst traffic bottlenecks in the country. Congestion on the Katy Freeway wastes millions of dollars annually due to lost productivity. For at least 11 hours every weekday, in both directions, congestion on the Katy Freeway reduces the mobility of local residents and adds significantly to the cost of doing business in the Houston area. This roadway was constructed from 1960 to 1968, and was designed to carry 79,000 vehicles per day. Thirty years later, more than 207,000 vehicles - of which approximately 22,000 are trucks - travel this stretch of highway every day. Through a major repaving and widening project, the Katy Freeway will be reconstructed for approximately 64.4 kilometers, from Houston's central business district to the Brazos River. This project, which starts in 2003 and will take ten years to complete, will cost an estimated one billion dollars.
The proposed reconstruction of the Katy Freeway is in a mature, developed urban setting and raises sensitive issues that will impact both property values and the lives of area residents. A number of large and small buildings will need to be modified, moved or demolished in order to accommodate this expansion. Properties in the public right-of-way will be assessed and acquired. TxDOT project development engineers have created design options via a schematic that considers traffic flow, safety, drainage, potential environmental impacts, and aesthetics.
In the past, only the schematics themselves were presented at public forums. TxDOT determined that the expansion of the Katy Freeway was so important and the issues so complex, that visual representation was needed to assist citizens in understanding the project and its component parts. After a review of existing data, TxDOT decided to purchase and apply LIDAR data for the development of visualization media associated with the I-10 project. Visualization Data Requirements/Decision Delivery Approach
The foundation for this project was the creation of a 3D model of some 12.9km of the proposed expansion project, derived from computer-aided design (CAD) schematics. Engineering applications, running in a CAD environment, were used to develop this complex 3D model. At the TxDOT headquarters in Austin, the Information Systems Division, Planimetrics Section, created a digital terrain model (DTM) of the existing freeway. Randomly selected breakpoints were identified from a computer comparison of scanned aerial photographs, using soft-copy photogrammetry to create the DTM. Two-dimensional (2D) cross section data were then created at regular intervals by using a commonly available roadway design software program. Interpolation between cross sections created the surfaces of the model. Figure 1 represents the very limited (e.g., width) DTM data derived from the hundreds of cross sections that were required to create the proposed project road surfaces.
To accurately display existing buildings and ground features, existing data are typically collected from several sources, including outdated U.S. Geological Survey (USGS) 1:24,000 map quadrangles. Available digital elevation models (DEMs) did not provide the existing features and topographic data accuracy needed for this project (Brown and Bara, 1994; Reutebuch and McGaughey, 1999). Existing data were also limited by TxDOT's standard width restriction, requiring them to be no wider than 182.9 meters from the project baseline. To be able to illustrate adjoining land and structures, the project needed digital cross sections of topographic features outside the 182.9-meter restriction.
Since TxDOT does not usually survey private property, nor does it have access to tall-building polygons and associated elevation data, TxDOT purchased a portion of the LIDAR data set. The initial data set represented all of Harris County and was created by TerraPoint LLC. This was the only available data set to provide a wide enough corridor area to suit the topographic data requirements of the project. The LIDAR data represented existing Earth features, with a laser hit posting of 3.0 meters and a vertical accuracy of at least a Root Mean Square Error (RMSE) of 0.4 meters.
From the LIDAR data set, a "bare earth" profile was derived. A bare earth data set - also referred to as a digital terrain model (DTM) - is an accurate representation of the terrain (Figure 2). The data set was developed using a semi-automated feature-extraction procedure. Large buildings, primarily commercial and more than 10 meters tall, were also located and mapped (Figure 3) by using the LIDAR data. The enormous amounts of information contained in this large data set were reduced relative to point density and imported into the CAD system. The reduced LIDAR information provided an accurate existing ground-surface topography with polygons representing adjacent buildings. All of this data was then incorporated into the 3D model.
Once the 3D model was complete, other visualization applications were then possible. For the expansion of the Katy Freeway, the 3D model provided two different media products. These were an engineer's review tool and a motorist's perspective of the future highway. The first provided TxDOT design engineers with a model that, when rendered, could be captured as images to be printed or displayed for review, analysis and editing. The second application created an animated "drive-through" of the proposed freeway. This second product represented a series of static images that, when viewed rapidly in sequence, gives a viewer the sensation of driving through the proposed project area. This series of images were then written to videotape and/or interactive CD-ROM and presented as public information. Roadway Engineering Design Review
Through experience, design engineers gain knowledge of how a proposed project will look once completed. The proposed reconstruction of I-10 was rated significant enough to create geometrically accurate visualization media for the engineering design staff to review and edit. The following list of tasks represents the visualization process, from the development of the initial road design schematic to the presentation of the proposed roadway to the public:
• Acquire existing cross section data and features
• Develop engineering design schematic
• Connect features such as frontage roads, ramps, main lanes, etc., on cross sections through interpolation to create a construction roadway surface
• Overlay proposed 3D model on LIDAR bare earth and add LIDAR building polygons, removing appropriate buildings within right-of-way
• Enhance 3D model with details such as concrete traffic barriers, guard rails, road signs, vegetation, pavement markings, etc.
• Map materials, colors, photographs and textures onto surfaces of 3D model
• Review engineering design of roadways and edit until approved
• Develop visualization media, such as rendered images for color printing, drive-through animation to be written to video, etc.
• Present visualization media at public forums for review and comment. Figure 4 represents a vantage point used by review and design engineers to assess a portion of the proposed roadway. This process included reviewing sight distances, conformity to national standards of safety, lane configuration, structural conformity, and possible aesthetic applications. Public Meeting Review
Public involvement is an essential phase of the highway planning process as authorized in the Intermodal Surface Transportation Efficiency Act (ISTEA, 1991), and in the Transportation Equity Act of the 21st Century (TEA-21, 1998), the federal laws that fund the national highway system. Interested persons are given the opportunity to become fully acquainted with proposed transportation improvements, both during the design process and prior to the submission of the final highway design for Federal Highway Administration (FHWA) approval (TxDOT, Highway Design Division Operations and Procedures Manual). This process assures the Department that supporting and opposing public views are expressed early enough in the planning process to allow for their proper consideration, while the design is still flexible enough to reflect these opinions. The results of public involvement contribute directly to the Department's decision-making process throughout the various stages of project planning and development. For the I-10 project, TxDOT conducted three initial public information meetings (November 8-10, 1999) to present the plans for the proposed roadway (TxDOT, 1999). During these meetings, input was received from community, business and governmental organizations. Comments and suggestions were evaluated by TxDOT for their ability to reduce congestion, improve safety, and reduce environmental impacts.
Before plans are approved, the Department is required to present the plans at a second, final series of public meetings. Final meetings for the expansion of the Katy Freeway have been scheduled for the third quarter of 2000. This will provide the public an opportunity to view and comment on the design of the proposed corridor, prior to its submission to TxDOT headquarters and the FHWA for approval. These meetings will be conducted in an open-house format that allows participants to review the schematic design - with highlighted proposed geometric changes - for the entire 64.4km freeway corridor.
The magnitude and complexity of the urban segment of the project (the I-610 loop west to Beltway 8) warranted the development of a 15-minute-long, animated drive-through for presentation at these public information meetings. The same visualization product used for the roadway engineering review is being used as the baseline product to develop the approximately 12.9km drive-through presentation. The public will be able to observe this drive-through to initially conceptualize and review the proposed roadway. Figure 5 represents a portion of the drive-through visualization product. LIDAR Technology and Data Processing
In the implementation of LIDAR technology (Flood and Gutelius, 1997; Schenk, 1999; Wehr and Lohr, 1999a; Wehr and Lohr, 1999b), a pulsed laser is directed out of the aircraft by a multi-faceted rotating mirror. Ground features intercept the laser pulse and reflect it back to the aircraft. The time interval between the laser pulse leaving the airplane and returning is measured accurately to derive the terrain. In post-mission processing, LIDAR time interval measurements are converted to distance and correlated to the aircraft's global positioning system (GPS), inertial measurement unit (IMU), and ground-based reference GPS stations. The GPS accurately determines the aircraft's position in terms of longitude, latitude and altitude. The IMU is used to determine the aircraft altitude in terms of pitch, yaw and roll. From these two data sets, the laser beam's exit geometry is derived relative to Earth coordinates.
Figure 6 represents the operational principles of the airborne and ground survey components for the LIDAR sensor system. For daytime flights, GPS time-stamped color video imagery or panchromatic digital camera imagery may be acquired in tandem with the laser data. Raw Georeferenced Data
Basic post-processing tasks, which create accurate x, y and z triplet data from raw LIDAR laser range, scan angle, IMU, and GPS data, are performed using proprietary post-processing code. Flight line side-lap data can be either included or omitted from the final product. After a complete quality assurance/quality control procedure, the LIDAR data are delivered in a geographic coordinate system - registered to the NAD83-199 adjustment datum - on CD-ROM in a variety of standard spatial data formats. USGS standard metadata files are also provided with each product. Bare Earth Process
A bare earth DTM is dependent, in part, on the type and density of vegetation (Bufton, et al., 1991; Haala and Brenner, 1998; Hoss, 1996; Kraus and Pfeifer, 1998; Pfeifer, et al., 1999; Ridgway and Minister, 1999). To generate a true representation of the topography, the laser beam must reach the ground. As the vegetation density increases, the error in a bare earth product is also likely increased. Conditions where there is less foliage (e.g., leaf-off conditions) provide the best density penetration. Proprietary software is used to perform the automated and semi-automated creation of a DTM. Using ArcInfo 7.0 functionality, the LIDAR-generated, bare earth results are based on the analysis of last-return laser pulses and statistical modeling. The software examines distance, angle of deflection, standard deviations, and other statistical tests for individual LIDAR data points, clusters of points, and associated neighborhoods. The user can introduce variables that recognize differing geomorphic, vegetation canopy, and human-built environments. Figure 2 represents a portion of the bare earth product along I-10. Tall Building Process
Depending upon project requirements, proprietary software is used either to semi-automatically or automatically construct vectorized tall (10 meters plus) building outlines, including elevations and sub-roof top geometry (Brunn and Weidner, 1997; Hug, 1997; Lemmens, et al., 1997; Maas and Vosselman, 1999). Figure 3 represents the results of the semi-automated data processing procedure that was used to configure tall buildings along I-10. Support pilings under the existing raised highway overpasses in Figure 3 were located approximately and inserted manually through the review of shadows on USGS and State of Texas-provided digital orthophoto quadrangles (DOQs). The DOQs were at a one-meter spatial resolution. Conclusions and Future Development
Use of the existing LIDAR data saved TxDOT approximately three months and the cost of a field survey crew, ancillary data gathering, manual vectorizing of the terrain and buildings, and associated personnel time. However, there were some aspects of the use of the project's LIDAR data that needed improvement. Engineering applications and CAD software used by TxDOT on this project had a data limitation of 32 megabytes, which could not accommodate the LIDAR data for the 12.9km segment of the project area. To solve this problem, TxDOT divided the project into three sections, reducing the point density of the LIDAR data in order to complete the project.
TxDOT's Houston-based Design Visualization Section staff discovered that commercial software, in general, must have its data limitations removed in order to keep pace with new, useful large-data sets such as high-resolution LIDAR and image data. TxDOT used the commercially provided, LIDAR-generated, value-added bare earth and tall-building polygon data sets. Once into the visualization project, TxDOT determined there were additional Earth features (e.g., marquees, billboards, etc.) whose representation would enhance the "drive-through" experience in the final visualization product. The process of adding these Earth features would typically require a survey to locate structures and measure their height, width and length. Therefore, TxDOT recommended the development of an interface that would allow for easy access, extraction and integration of the raw LIDAR data into the existing TxDOT engineering-applications development software. Summary
State Departments of Transportation (DOTs) play a vital role in keeping the nation's highways maintained and traffic moving along. LIDAR technology provides an efficient and cost-effective solution to the need for accurate digital elevation data. Decision-making can be stronger when engineering and visualization application products are used in tandem. Visualization techniques present geometrically accurate facts for engineers and the public to view, assess, comment and edit proposed highway projects. In the future, DOTs will use 3D visualization capabilities that incorporate the most advanced spatial data, geographic information system (GIS) technologies, and remote sensing technologies to review, analyze, edit and approve the majority of projects. TxDOT envisions LIDAR data to have the potential use for planning activities, hydrology/drainage studies, litigation support, environmental analysis/permitting, subsidence/maintenance planning, and near-real-time emergency response/traffic routing. Acknowledgements
The authors are grateful to their colleagues at TxDOT's Houston District, the Houston Advanced Research Center (HARC), and TerraPoint LLC for the LIDAR data processing and products presented in this article. Particular appreciation is given to consultant Sheldon Post for his assistance in the development of the final products. About the Authors:
John M. Hill is with the Environmental Information Systems Laboratory, Houston Advanced Research Center (HARC), The Woodlands, Texas. His e-mail address is [email protected].
Cindy Gloyna and Kathleen Chavez are with the Houston District's Information Resource System Section of the Texas Department of Transportation (TxDOT), Houston, Texas. Their e-mail addresses are [email protected] and [email protected] respectively.
Dana Young is with TerraPoint LLC, The Woodlands, Texas. E-mail address is [email protected]. Back |
