LIDAR and Updating Base Maps

Karen F. Adkins

Maximizing existing base map information in the process of updating GIS databases has great importance for GIS users. However, the accuracy and quality of some existing vector data such as contours may be insufficient as to warrant replacement. Light Detection and Ranging (LIDAR) technology provides a source of highly accurate new digital elevation/terrain models (DEM/DTM) for combination update/new base map development projects. This paper examines the advantages of this approach in update projects, identifies obstacles and observances of the process, and considers the future of LIDAR as a tool in update mapping.

Updating existing base map data typically involves decisions about what data needs to be updated, what needs to be retained, and whether or not to change datum and/or accuracy requirement of the base map. The intensity of the decision making process varies with the base data to be updated. Digital orthophotography is often periodically updated. It basically involves photography of the areas requiring update, and a process of rectifying new imagery using control and digital elevation or terrain model (DEM/DTM) from the existing base map, replacing only changed DEM that affects the accuracy of the imagery. Updating vector data involves a more meticulous process of identifying change in the areas to update. This may involve a detailed model-by-model review of the base data in conjunction with the old imagery. Sometimes the datum of the base map must be updated, requiring an even more intensive process of transformation. Global and model-by-model transformation of the new mapping to the old are applied in such cases. Update techniques vary, but all result in accomplishing the goal of modernizing the base data.

Techniques and Management of Update Projects

Existing data must be imported into the production environment, often a stereocompilation workstation format such as Microstation. Pre-processing can also involve draping two-dimensional features whose original elevation is was lost or discarded onto a three-dimensional DTM/DEM to accurately detail changed features. This enables the compiler to view and update the features in stereo, so features appear at or near the proper elevation relative to the imagery. Care must be taken to effectively retain indispensable existing attribution in the database. Old and new features can also be tracked with attribution, differentiating the mapping cycles with information such as date of photography. When LIDAR is introduced into the update project, the intensive task of update of the DTM is alleviated. The LIDAR DTM also provides a highly accurate base for draping existing features.

LIDAR is universally used as a tool to measure change. The use of LIDAR data for disaster management purposes, such as recovery of the World Trade Center and hurricane destruction analysis, have sparked trends in applying LIDAR for update projects in general. Many industries are tapping into the LIDAR potential; for example, changes in coastal topography and forest canopies are some of the more commonly measured subjects. Comparison of data from follow-up flights assists in detection of change, and provides an understanding of long-term trends (NOAA, 2002). The Federal Emergency Management Agency (FEMA) has defined standards for and is modernizing floodplain maps throughout the United States using LIDAR data in its Map Modernization Initiative. With the prominence of such programs, using LIDAR as an update tool is steadily gaining support in the development of large-scale mapping programs. This has subsequently led to the use of LIDAR data in updating existing base maps. In particularly fast-growing communities, LIDAR is a very effective method of updating large areas of change by providing new DEM/DTM. It is especially effective in providing a new base DEM/DTM for orthoimagery and contouring, when global replacement of such features is warranted. In many cases, LIDAR DTM can assist in achieving better results in the update process. But it is important to keep in mind the same variances between old and new data will exist in LIDAR update projects as in traditional update projects.

The initial planning of the update project requires outlining specific guidelines for the mapping process. What accuracy standards must be met; what final purpose will the data serve; what is to be retained, removed, and changed - these are all critical steps in finalizing the update process. In projects using LIDAR, the expectation is often that the DTM collected will provide a higher-accuracy base with which to update the existing data. For new orthoimagery and topography, this is usually true. However, when incorporating other vector data into the process, perceptual and technical issues are often exaggerated in the base data. It is crucial to the success of the mapping process to determine in advance where the line is to be drawn between old and new mapping.

Careful planning for this process is necessary, as combining existing data with new data introduces both technical and perceptual problems from interpreted variations in old and new compilation. Variations between data mapped at different periods result in “fault lines” along the boundaries between the data. And, while meeting the target accuracy, the data may appear to vary significantly with the new imagery. In fact, both the LIDAR data and compiled features may be within the tolerance, but appear not to agree with each other (T. Greening, personal communication Jan. 4 2002). However, accuracy improvement is achievable when effectively using the proper techniques and taking into consideration the independent mapping of the old and new data. Refinement in the RMSE checks on models should be made between old and new mapping to prove updated maps meet specific accuracy standards (A. Thorpe, personal communication Jan. 4 2002).

Awareness of problems encountered on past LIDAR update projects can help future projects flow more smoothly through the production cycle. In the last several years, combination LIDAR/update projects have educated both developers and users in the planning and management of these types of projects. The following case studies provide examples of such technical and managerial techniques and observations.

Case Studies

The City of Edmond project consisted of replacement LIDAR DTM for generating new orthos and 2’ contours as well as update of the City’s planimetric data. In the case of Edmond, both accuracy and interpretive issues were encountered in the stereocompilation update process. The existing base data was compiled from 1”=800’ photography, to ASPRS Class I Accuracy Standards for 1”=200’map scale. The current mapping was performed using 1”=600’ photography; other specifications remained the same as the existing data. With this variance in scales and achievable accuracy, stereo comparison of the old data with the new LIDAR DTM accentuated the dramatic differences between the existing and new mapping. Features that had not been updated for over 6 years presented varying degrees of questionable compilation, at times interfering with what should have been straightforward capture of new features. Hydrology was a prime example of problematic base map data, and in most cases was replaced completely.

Another combination update/new mapping project, the City of Round Rock TX, identified fewer inaccuracies between the old and new data but hosted other management problems. Time lapse of greater than 3 months between the acquisition of the photography and the LIDAR data occurred, due to unfavorable weather conditions. The high-growth satellite community of Austin presented LIDAR data that did not agree with the new photography, particularly in areas where the Interstate 35 corridor was experiencing rapid and substantial change. And while the existing planimetric data was of a somewhat better quality than what was encountered in the Edmond project, resolving the draped data with the photography was time consuming and at times difficult to perform. The compilers often identified the LIDAR data as in error. For example, large fields appeared to have “floating” points, as did several areas under construction. In these instances, the features causing the apparent incorrect elevations had been removed during the time gap.

Planning for the Future

Update projects vary somewhat in scenario, which to some degree will depend on the methodology of the update approach. Using a LIDAR methodology can be very beneficial in combination replacement/update projects, but proper planning of traditional update issues that may be compounded by the use of LIDAR will help future update projects succeed.

LIDAR data is easily viewed and manipulated within the Esri suite of tools. Tools such as the LIDAR Loader Extension and the LIDAR Data Retrieval Tool (LDART) allow for easy acquisition and rapid import of LIDAR data sets directly into ArcView. Yet Esri is currently challenged to provide production tools that will allow geodatabases in seamless, versioned environments to be imported into a compilation environment and exported back to the same versioned environment. Update projects often rely on tracking changed or modified features and attributes, and the current Esri tools function to enable the end user easy maintenance and tracking through versioning. Scripts developed in alternative platforms such as Oracle and SmallWorld allow data that has been tracked in compilation as new, modified or deleted to be imported through tables to various user platforms including versioned geodatabases. This process is currently implemented for all 5 boroughs of the New York City update project.

Conclusion

LIDAR has become a fixture of present-day mapping missions, providing cost-effective means to achieve high-accuracy results. LIDAR lends itself as a tool to make more frequent updates possible for high-change areas. Continued improvements in the technology of the equipment and processing tools include better accuracy and feature extraction, which will assist in more effective update of base maps. GIS managers aim to achieve greater accuracy of data over time through periodic updates, while retaining valuable database information. Development of tools within platforms such as Esri to assist with the update process will undoubtedly add to the value of using LIDAR to accomplish such goals.

References

References Brock, John C. et al., 2001. “Recognition of Fiducial Surfaces in Lidar Surveys of Coastal Topography.” Photogrammetric Engineering and Remote Sensing 67 (11): 1245-1258.

Chapman, Bruce. “SAR Interferometry and Surface Change Detection.” Hydrology, Ecology, Environmental Monitoring and Global Change. 1995. http://southport.jpl.nasa.gov/scienceapps/dixon/report5.html.

Daniels, Robert C., 2001. “Datum Conversion Issues with LIDAR Spot Elevation Data.” Photogrammetric Engineering and Remote Sensing 67 (6): 735-740.

Dubayah, Ralph O. and Jason B. Drake, 2000 “LIDAR Remote Sensing for Forestry.” Journal of Forestry 98 (6): 44-46.

Federal Emergency Management Agency. “Map Modernization Update: FEMA Uses New Mapping Technology.” December 7, 1999. http://www.fema.gov/nwz99/lidar1207.htm. (May 9, 2002).

National Oceanic and Atmospheric Administration. “Topographic Change Mapping.” March 29, 2001. http://www.csc.noaa.gov/crs/tcm. (June 3, 2002).

National Oceanic and Atmospheric Administration. “Using LIDAR in an ArcView Project.” http://www.csc.noaa.gov/products/sccoasts/html/dataload.htm. (June 3, 2002).

Thomas, T. J. Personal Interview. June 6, 2002.

Turner, Keith A. “LIDAR Provides Better DEM Data,” 2002. http://www.geoplace.com/gw/2000/1100/1100/agf.asp. (June 3, 2002).

I would also like to thank Trevor Greening, Krysia Sapeta, and Tim Thoma, for providing a wealth of information about update mapping and LIDAR.

Karen F. Adkins
Project Manager
Sanborn