The National Cooperative Geologic Mapping team (NCGM) in the Central Region of the U.S. Geological Survey (USGS) has some unique capabilities for collecting geologic data from old (printed) geologic maps, gathering information for new maps, and updating of existing digital maps. Maps on frosted or clear mylar (and in rare cases on paper) are scanned and vectorized almost automatically. Data gathered by both vectorization of scanned maps and by photogrammetric techniques from aerial and oblique photography are then imported into Arc/Info. Geologic map coverages are created and maintained in Arc/Info, and updates and corrections are made in Arc/Info from data captured by these techniques. Scanned mapping allows the measurement of line placement accuracy from pre-existing map materials. This measure may be totally unrelated to original line placement by a geologist in the field. Photogrammetric analysis of oblique photography of the land surface allows the capture of true three-dimensional geologic data. This technique can be especially useful in areas of rugged terrain and locations that are difficult to reach, or in field conditions where studies are possible only for a short time. Photogrammetric monitoring of volcanic activity is an ongoing research problem.
The USGS has produced digital maps for many years, but the earliest of these were representations of geophysical data or remotely sensed data. Traditionally, most geologic maps were drawn by hand from observations made on aerial photographs or directly on topographic map bases. These maps were first turned into a digital state when scribe coats or film positives of the line work and symbols were scanned with National Mapping Division's SCITEX equipment. Line work was "cleaned up" and correct symbols were added to make higher quality printed maps. The next step in vector data capture to make digital geologic maps for the Geologic Division scientists came with the USGS-developed software package called GSMAP. The last version of this software (Selner and Taylor, 1993), running on standard PC hardware allowed the geoscientist to digitize old and new geologic mapping and transfer the data set to the SCITEX environment or to Arc/Info, GRASS, or AutoCAD.
In 1972, the Central Regional Geologist's office established a plotter lab and funded the acquisition of the analog Kern PG-2 plotters. These devices allow the transfer of geologic information from stereo aerial photography to registered green line quadrangle or paper "ozalid" maps.
In the early 1990s, digital encoders were added to two of the PG-2 plotters and software was written to allow digital capture of the line work from the aerial photography and transfer of the resulting data into AutoCAD, GRASS, and Arc/Info. At roughly the same time a Kern analytic plotter was acquired by project activities sponsored by the U.S. Department of Energy (DOE) working at Yucca Mountain near the Nevada Test Site, Nevada. This instrument, depending on the quality of the photography, can capture quite usable data at a resolution of 3-4 microns on the photograph. The DOE supported this high-precision system in order to map geology over small areas in tunnel walls at Yucca Mountain (only briefly exposed during tunneling) and in trenches excavated for specific scientific studies.
Most recently a "Soft Plotter" hardware-software system from Zeiss, Inc. was added to the plotter laboratory. This system will allow for more photographic data to be used for the rapid collection of geologic data from many different image data sets. In one use, scanned aerial photography captured at a resolution that is a multiple of seven microns on the photograph (on Zeiss scanners) will allow as good a correction to be applied to rectify the photography to the ground as could be done using an analog plotter. The overall accuracy is probably better because hundreds of control points can be located on the photographs for a better approximation of the corrections necessary to register the photography to the ground.
Data arrives in the Data Acquisition Facility (DAF) in a number of ways. First, a geologist draws their line work representing their field observations on aerial photographs (stereo pairs) of the area in which they are working. For a 7 1/2' quadrangle this may mean 25 stereo pairs on average to cover the entire quadrangle. Some geologists record information directly onto the paper topographic sheets. These data are later digitized using a number of software packages including GSMCAD, AutoCAD, DesignCAD, and Arc/Info.
Pre-existing maps, both published and unpublished are additional sources of geologic data. To use the current scanner to its best advantage, these data must be either on a film positive (the best way) or on translucent mylar. Paper maps are usually not acceptable for scanning. For published maps, some of the original publication archive materials still exist and is suitable for scanning. For other maps, data must be traced onto mylar from the paper copies, a tedious procedure. Some data will probably be misrepresented in this process. Tracing geology works better when the data are generalized and compiled at a smaller scale.
In addition, the USGS and other civilian agencies have access to image data from National Technical Information sources. Data derived from these sources are usable on the "Soft Plotter" system.
Scanned vector data from the USGS Central Region GIS Facility is turned into vector data by passing through the software package LT4X. It has been found that LT4X has more features that are quite useful and efficient to "vectorize" raster data than other commercial software packages that were investigated. First, the software computes a minimum horizontal accuracy with which lines can be located. For a 1:100,000-scale map that is scanned at 400 pixels/inch, the resolution is 15 meters. Another feature is a dynamic calculation of the accuracy with which the scan is geo-referenced. The coordinates of the map are produced in a number of map projections. In addition, when the appropriate zone numbers are furnished, conversions to UTM and State Plane grids are computed as well.
Before vectorization in LT4X is done, LT4X first reduces scans of individual lines to the width of one pixel, then spurs on lines are removed, holes in lines are filled and vectorization proceeds. After vectorization, the vectors can be checked against the original scan to allow differences between the vectorized map and the original scan to be corrected. Polygon closure is automatically checked for improper polygons, and when polygons are attributed, adjacent polygons with the same attribute set are noted to allow for correction if needed. When one is needed, a geographic quadrangle boundary in raster mode can be added before vectorization to allow for closure of polygons at the edge of a map.
Geologic line work on stereo aerial photography can be collected in digital as well as analog form in the Photogrammetry laboratory of the DAF. In the standard photogrammetric procedure in use, the laboratory photogrammetrist, Jim Messerich, "sets" models on a PG-2 plotter using the field stereo pairs. Then for the digital PG-2 plotters, the line work is traced by the geologist and subsequently is transferred to a file on a Data General workstation using CADMAP, a software package from Zeiss, Inc. These data are displayed on a CRT screen as they are captured on top of hypsography or hydrography (when these are available) to act as a check on the registration of the geology and to insure that the line work "fits" well with the various data layers on a topographic quadrangle.
When an analog PG-2 is used, geologic data are transferred directly to a mylar sheet using a pantograph system. Usually the mylar has green-line drawing of the topographic quadrangle on it. Newer techniques use a punch-registered mylar sheet on top of the green line sheet. Data generated in this fashion are hand digitized, or scanned for use with the raster-to-vector process described above.
The "Soft Plotter" hardware/software subsystem is to be installed in the near future and should be of great utility when it is done. This system allows data to be collected from digitized photography or other appropriate imagery. The system also permits photogrammetric models to be more easily and quickly set. In addition to the capture of geologic data in a timely manner, the geologist can have a topographic contour map created dynamically to check the location of point and line data. When three-dimensional topography is generated and geologic data is collected on this surface, one additional problem must be solved. Differences occur between published topography and the topography generated by this technique. The proper placement of geologic data on one or the other of these two representations of the topography is difficult.
The resources needed to store high-resolution scans of aerial photography are substantial. One 9" x 9" photograph scanned at a resolution of 7 microns on the photograph and 256 gray levels occupies about 55 Mbytes of disk storage.
The photogrammetric laboratory has additional capabilities to collect geologic data from oblique photography shot by tripod-mounted hand-held cameras. The lab has a camera calibration room for such cameras that allow the measure of the correction factors for used camera-lens combinations. These corrections allow the use of photography taken with these cameras to be used to obtain three-dimensional geologic data. Photography generated this way is called oblique stereo photography. An explanation of how this photography is used will be given in the examples below.
The Digital Geologic Map of the Nevada Test Site (Wahl and others, 1997) is an example of a product that used multiple techniques to obtain vector data from raster scans or from photogrammetric plotters. Final editing of the vector data was done in Arc/Info, but seven of the nine coverages plotted on the map were initially captured using scans of existing map data or aerial photography on photogrammetric plotters. All the plotter data were captured digitally. The surficial geologic data in Yucca Flat and Gold Flat areas on the Pahute Mesa 1:100,000 part of the map were added into the database using LT4X and data from scanned film positives. In the Springdale 7.5' quadrangle, in the south central part of the Pahute Mesa 1:100,000 sheet, data from the digital PG-2 were made into an Arc/Info coverage, and added to the map database from Arc/Info. The background topographic and planimetric map data were created from scans of the film positives of each 1:100,000 sheet and vectorized after the data were geo-referenced in.
Three recent projects demonstrate the capabilities of the photogrammetric laboratory. The first project used a before-and-after study of the aerial photography of the Landers, California 1994 earthquake to document ground deformation patterns in detail. Photography at scales of 1:2500 and 1:6000 along with the new field control points allowed the motion along the fracture zones associated with the earthquake to be measured from the photography with an accuracy of from 2 cm to 3 cm. This study is in a research stage still and is being done by the Central Region Earthquake and Volcanic Hazards Team (EVH) and the photogrammetry laboratory. The collaborators are Robert Fleming, EVH, and Jim Messerich, NCGMP.
Work by Ren A. Thompson on the chronostratigraphy and eruptive history of this Chilean stratovolcano was successfully completed because the photogrammetric laboratory has the capability to capture geologic data from oblique photography (Singer and others, 1997). Because the rugged terrain would call for much time to gather three-dimensional geologic data (with less accuracy), tripod-mounted 35-mm cameras were used to make overlapping photographs of the canyon wall cross-sections through the volcanic edifice. The analytic stereo plotter was used to obtain not only horizontal positions but also altitudes at the same points. This means that true three-dimensional geology can be captured by this technique and is more timely and more accurate. This contrasts with the accuracy of vertical coordinates for measured horizontal positions derived from the interpolation of contour lines in a topographic base map.
Oblique photogrammetry was used to measure the extent of decay from atmospheric pollution on the marble of the columns on the outside of the Merchants Exchange Building in Philadelphia, Pennsylvania. This was work was done for the National Park Service. When the photography was analyzed, the resolution of the plotter gave measurements accurate to within one millimeter. This is another situation in which oblique stereo photography, when analyzed with high accuracy photogrammetric plotters, can yield better accuracy and faster results than conventional methods in the same situation.
The volcanoes of the Aleutian Islands, particularly St. Augustine Island have been observed by oblique photography when helicopter availablility and weather conditions permitted. This was done as part of an experimental program to remotely measure ground surface motion on volcanoes. Now other stereo imagery could be used in the same manner. Control points were set and located for base locations. With imagery other than photography the imagery must have sufficient resolution to resolve such ground control points. As the surface of the island moves in response to volcanic activity, the size of the motion can be accurately measured from the imagery.
Geologic data from old or new geology can be captured from scanned line work drawn on film positives or mylar sheets with the currently-owned scanner. These data can be quickly geo-referenced and converted to vector data suitable for use with CAD software or GIS software like Arc/Info.
Photogrammetry can be used to quite accurately and quickly capture geologic data drawn on aerial photographs that would be suitable for use with CAD or GIS software.
Photogrammetry can be used to capture unique data sets and provide insights into geologic processes. In the case of the Landers, 1994 earthquake photogrammetry provided an excellent way to document with great precision ground deformation associated with an earthquake fault zone rupture pattern.
Geologic data can be accurately collected from near-vertical exposures; assembled and analyzed in three-dimensions to precisely describe rock unit relationships especially in rugged terrain.
The capture of three-dimensional geologic data from stereo imagery provides a level of geologic information that GIS systems cannot handle yet, because all major GIS software systems store topology in two, not three, dimensions.
Selner, Gary I., and Taylor, Richard B., 1993, System 9, GSMAP, and other programs for the IBM PC and compatible microcomputers, to assist workers in the earth sciences: U.S. Geological Survey Open-File Report 93-511, 363 p.
Singer, B.S., Thompson, R.A., Dungan, M.A., Feeley, T.C., Nelson, S.T., Pickens, J.C., Brown, L.L., Wulff, A.W., Davidson, J.P., and Metzger, J., 1997, Volcanism and erosion during the past 930k.y. at the Tatara-SanPedro complex, Chilean Andes: Geological Society of America Bulletin, v. 109, no. 2, p. 127-142.
Wahl, Ronald R., Sawyer, David A., Minor, Scott A., Carr, Michael D., Cole, James C., Swadley, WC, Laczniak, Randall J., Warren, Richard G., Green, Katryn S., and Engle, Colin M., 1997, Digital geologic map of the Nevada Test Site Area, Nevada: U.S. Geological Survey Open-File Report 97-140 50p.