Orthophotos combine the image characteristics of a photograph with the geometric qualities of a map. The original CIR-DOQQs mapped a 1-meter ground resolution of a DOQQ (a 3.75-minutes of latitude by 3.75-minutes of longitude extent) plus the overedge ranging from about 50 meters to 300 meters. The overedge is included to facilitate tonal matching for mosaicking The normal orientation of data is by lines (rows) and samples (columns). Each line contains a series of pixels ordered from west to east with the order of lines fom north to south. The DOQQs are an 8-bit composite raster with radiometric image brightness values stored as 256 levels, from 0 to 255.
The above link also has the file(s) DOQQdb.htm and/or DOQQdb.txt; the metadata of the above database.
To download the DOQQs, visit <http://dnrweb.dnr.state.md.us/gis/data>
Slight systematic radiometric differences can be detected between adjacent DOQQ files due primarily to differences in source photography capture dates and sun angles of aerial photography along flight lines. These differences can be observed in an image's general lightness or darkness when compared to adjacent DOQQ file coverages. These DOOQs were produced to meet the National Map Accuracy Standard for 1:12000 scale.
EXPLANATION OF GROUND CONTROL:
The DOQQs were produced in batches. Each batch utilized various ground control methods. The RMSE values are for each batch and not the individual DOQQs. The batch areas are classified by the Project_ Area. The Project_ Area information is available from the following link:
Downloading the DOQQ data from <http://dnrweb.dnr.state.md.us/gis/data/> yields the file, DOQQdb.csv. This file has the all important Project_Area information for any given DOQQ.
Following provides the explanation of the various Ground Controls used.
A. Obtained from 4 sources:
1) Aerial spot photographs at 1:12,000 scale photography which showed panel locations of existing USGS/NGS control.
2) Previous USGS aerial triangulation of part of Delaware which overlapped into Maryland.
3) Previous USGS aerial triangulation of part of Virginia which overlapped into Maryland.
4) USGS 7.5' Quad maps to obtain additional vertical ground control.
B. Obtained from 3 sources:
1) Aerial spot photographs at 1:12k scale photography showing panel locations of existing USGS/NGS control.
2) Previous USGS aerial triangulation of part of Virginia which overlapped into Maryland.
3) USGS 7.5' Quad maps to obtain additional vertical ground control.
C. Aerial target panels placed on existing monuments with field surveys performed using conventional methods for offset panels.
D. GPS control of aerial target panel locations - monuments may or may not have been set.
E. GPS control of photoidentified features.
F. Airborne GPS.
RMSE VALUES:
Ground control method(s) used and RMSE values, in meters, for various project areas are:
Project Area 1; Ground Control used is 'A'. RMSE in X direction is 3.066, in Y direction is 2.432, in Z direction is 0.611
Project Area 2; Ground Control used is 'C'. RMSE in X direction is 1.750, in Y is direction 1.508, in Z direction is 0.652
Project Area 3; Ground Control used is 'A'. RMSE in X direction 3.939, in Y direction is 3.950, in Z direction is 1.004
Project Areas 4 & 7; Ground Control used is 'D'**. RMSE in X direction is 2.300, in Y direction is 1.800, in Z direction is 0.628
Project Areas 5, 12 & 13; Ground Control used is 'B'**. RMSE in X direction is 3.074, in Y direction is 3.048, in Z direction is 0.715
Project Area 6; Ground Control used is 'C'. RMSE in X direction is 1.897, in Y direction is 1.691, in Z direction is 0.853
Project Areas 8 & 9; Ground Control used is 'D'. RMSE in X direction is 0.975, in Y direction is 0.772, in Z direction is 0.755
Project Areas 10, 11,14 &15; Ground Control used is 'C'. RMSE in X direction is 1.977, in Y direction is 1.316, in Z direction is 1.532
Project Areas 16, 17,18 &19; Ground Control used is 'E & F'. RMSE in X direction is 1.823, in Y direction is 1.616, in Z is direction 1.341
Project Areas 20, 21 & 22; Ground Control used is 'E & F'. RMSE in X direction is 1.823, in Y direction is 1.616, in Z direction is 1.341
Project Areas 23, 24, 25, 26 & 27; Ground Control used is 'E & F'. RMSE in X direction is 0.280, in Y direction is 0.202, in Z direction is 0.812
Project Area 28; Ground Control used is 'E & F'. RMSE in X direction is 0.568, in Y direction is 0.715, in Z direction is 0.515
Project Areas 29 & 30; Ground Control used is 'E & F'. RMSE in X direction is 1.193, in Y direction is 0.688, in Z direction is 0.489
Project Areas 31, 32, 33 & 34 contain areas of only water or are outside Maryland. These were not produced.
** Aerotriangulation was performed in separate blocks. The RMSE values represent the greatest error for any block within the project area. Again. The RMSE values are in meters.
INTRODUCTION
There is an intense national interest in digital mapping programs conducted by all levels of government. Sharing of technical studies and knowledge is vital to prevent duplication of effort and to accelerate the GIS activities and capabilities of all agencies.
The Maryland Department of Natural Resources (MD DNR) maintains a modern mapping program that combines innovative technologies with established mapping principals. DNR produces digital orthophoto maps that exceed National Map Accuracy Standards at 1:12,000 scale. These maps provide the base for a high quality wetland inventory and are produced at an average cost when compared to other programs. They have the advantages of being flexible and easy to update due to their digital design, and they are easier for the public to use due to the image base.
EXPLANATION OF MARYLAND'S DOQQ MAP PRODUCTION
Production of DOQQ map products can be accomplished in a relatively short period of time and provides the widest possible range of output products. The Maryland DOQQ maps are the ideal base layer for geographic information systems (GIS) which can use them at scales smaller than 1:3,000. High quality photographic, paper and mylar products have been created at scales ranging from 1:4,800 to 1:100,000. Useful hardcopy (printed) products have been created at larger scales, but generally do not have good edge definitions (fuzzy edges).
DNR has maintained an active role in decisions regarding production of these maps and arranged for control and paneling of monuments until the advent of airborne GPS control. MD DNR also performs quality assurance of the production products. These efforts have helped to keep the costs as low as possible to the State. This program has been a good example of a cooperative working arrangement between local and State government agencies.
PROJECT AREA
Map production is based on complete coverage of individual project area counties. DNR has formed partnerships with several counties to provide a sharing of expenses for a product that meets the needs of both levels of government. Our production has been guided by the following factors:
Concern for areas that are experiencing significant development pressure.
Areas that are scheduled to have soils remapped by the Soil
Conservation Service.
An initiative to study the coastal bays on the Atlantic Coast.
Counties that are willing to fund a portion of production costs.
Specific project areas (e.g. watershed studies).
Counties that are most likely to use the digital data in a GIS system application.
After weighing the above factors, we work during the summer to determine our production schedule for the following year. We meet with interested county agencies to determine their level of participation. Several counties have funded a substantial network of control monuments and aerial photography that is more current than that available from the NAPP program.
These arrangements are firmed-up by October and a letter of intent, or purchase order, is entered into by the State and county. We typically enter into one or more meetings with the county to ensure an understanding of the respective roles, and the products that will be exchanged.
OBTAINING A SUITABLE CONTROL NETWORK
Maryland's initial specifications required that control be provided at 3.75 minute intervals of longitude and latitude, roughly at the corner of each quarter quad map. MD DNR contacted other State and local agencies to determine if suitable control monuments exist at the appropriate intervals. After making this determination, we meet with our contractor to locate sites where additional control should be provided to augment the existing monuments. Those sites are surveyed by cooperating counties or by one of the State agencies with survey capabilities. We typically prefer to use GPS in these situations. During the first year of production, we found that we could spend an entire day traversing a survey to a single point in rural areas. Only one of the first eight counties had an adequate network of monuments. Three counties (Queen Anne's, St. Mary's, and Frederick) opted to provide complete coverage using an entirely new GPS network. Full county coverage normally requires 60 to 80 control points. This work was typically completed by January of the year in which the maps were produced.
MD DNR obtained pugged photographs and the aerotriangulation report from USGS for portions of the first project area involving Worcester, Wicomico and Somerset Counties. This was done under the assumption that it would reduce costs over a relatively large project area. The control network was created for 1:24,000 scale map production and we found that USGS made assumptions on tide height at the shoreline of islands for vertical control that was unacceptable for 1:12,000 scale maps. We provided approximately 110 additional spot photographs of new control points to increase the density of the existing network. A significant amount of survey time was spent by State crews to obtain the necessary data. We also had to fly stereo coverage over known State boundary markers in open water areas (large pilings) to successfully control the photography over large open water areas. In retrospect, this effort was probably as expensive as a new GPS network in terms of labor costs, and was certainly more difficult to orchestrate. Existing monuments are frequently found in areas that are not ideal for aerial photography. In these cases, new GPS points can easily be located on grass at public properties.
Horizontal and vertical control is provided to third order or higher precision. Vertical control is provided at less frequent intervals than horizontal control. Control is collected in feet using the Maryland State Plane Grid Coordinate System with NAD 1983 for horizontal points and NGVD 1929 for vertical points. Particular attention is given to the number of decimal places used in conversion programs. Maryland is over 1,800,000 feet in length, (340 miles), and if precise data conversion is not performed, it will impact the ortho correction process and produce maps with unacceptable edge matching.
Beginning in 1994 DNR elected to use airborne GPS control. GPS data is collected for each aerial photograph during the flight. This provides for control at 3.75' intervals. Additional control is provided by using GPS survey equipment at photo identifiable features spaced at appropriate intervals across the project area to tie the airborne GPS to the ground. This eliminates the need to panel existing control monuments and maintain the panels until the photography is acquired.
PANELING CONTROL MONUMENTS
Prior to aerial photography, panels (a.k.a. aerial targets) will be placed over the control points. This work is typically completed in early to mid March. The panels are T-shaped with each leg being 16 feet in length. The material is a 24-inch white plastic that is attached to the ground using cardboard reinforcements along the outside edges that are pierced with a nail and metal washer. The intersection of the T is centered over the control monument.
Typically a team of two personnel will panel up to 15 control points in one day. The teams are comprised of employees from the county and/or State. It is often very difficult to locate the existing survey monuments in counties that do not use witness posts, even though we have a description of the site and measurements from landmark features. It is frequently necessary to take metal detectors to locate the control disks. Teams without experienced surveyors may only panel three to four (3-4) sites in one day.
In the event that the selected control monument can not be found, or is not visible through obstacles such as heavy brush, trees or highway overpasses, an alternate monument may be selected. Frequently, this is not feasible and a panel is located in a suitable public location that can later be surveyed. We drive a piece of steel rod at these locations that can be located if the panel is destroyed. We record measurements from other features and document the locations.
In many cases, the existing control monuments are in private yards or posted properties. If the property is posted, we do not enter. In other situations, we approach the residents and explain the program in hopes that they will permit us to place the panels. When property owners are not home or it appears they are not actively using the property, we place the panel and leave a letter explaining the program that has a contact name and phone number. Most private property owners are cooperative, but some will refuse permission. Others will simply take the panel up without contacting us.
The panels must be inspected every few days to ensure that they are in good condition. They are frequently damaged by wind and human or animal activities. Repairs are made as required.
AERIAL PHOTOGRAPHY
Maryland requires the use of USGS National Aerial Photographic Program (NAPP) specification photography for production of its DOQQ maps. There is a significant federal specification manual for this photography. It was not necessary for the State to develop its own specifications. This photography is flown at approximately 20,000 feet (+ or - 5%) using an aerial camera with a six-inch focal length lens that produces 1:40,000 scale photography. Maryland requires a color infrared emulsion that is obtained in early spring with a leaf-off condition. This significantly improves wetland delineations when compared to black and white emulsions or leaf-on conditions. Leaf-off conditions also make the collection of elevation data possible. Maryland has produced maps for three counties (Wicomico, Worcester and Somerset) using NAPP photography that is available off-the-shelf from USGS. The remaining counties have been mapped using newly acquired photography that meets NAPP specifications. A very tight window of opportunity exists to acquire the photography. We can readily demonstrate that photography acquired after lawns and winter wheat have greened-up will produce a much more attractive image base. If photography is acquired too early in the spring, the image base will appear dull or flat. The issue is the amount of contrast in the infrared band that depicts active vegetation. Typically there is a 4-5 week window for acquisition of photography and we are very lucky if we get 4 days that are suitable.
Atmospheric conditions must be ideal for this photography. The perfect scenario for wetlands mapping is a moderate rain followed by a cold front with associated winds to dry the upland areas. This must be followed by a day with no clouds or haze, during which the contractor takes the photography.
Three diapositive prints of each photograph are made. One is used by the photo interpreter and maintained by the State. One is provided to counties that pay for photography. The last is maintained by the contractor for map production. In situations where we do not fly new photography, off-the-shelf diapositives are ordered from the USGS EROS Data Center in Sioux Falls, Iowa.
PUGGING CONTROL
The contractor uses a very fine drill and places a hole in the photographs at the intersection of each panel or GPS point. These points can be identified later during the ortho correction process. For Worcester, Wicomico and Somerset counties, panels were not placed when the NAPP photography was acquired. For this case, the contractor aligned features in the spot photography with the NAPP photography and cross-pugged the point between photographs. In some circumstances, no panels are present in any photography. When this occurs, the contractor identifies features in the photograph that can be precisely located on the ground. The contractor provided personnel to survey these points and pugs the points using the field data.
AEROTRIANGULATION SOLUTION
The contractor performs an analytical aerotriangulation solution on the photography. This is a complex process of mathematical computations that compares the known control points with each other and with photo identifiable points common to adjoining photographs. Each point is compared in a triangulated computation with its neighbors. Data on the camera lens that is taken from a camera calibration report is also factored into this equation. The result is a list of Root Mean Square Errors (RMSE) that are also known as residuals. They predict the measured deviation of each point on the photograph from the theoretical computation. These residuals are compared to a known maximum value to ensure that the maps will have the correct geometry, thereby providing adherence to National Map Accuracy Standards (NMAS). The contractor delivers a copy of the aerotriangulation report which is retained by MD DNR.
DIGITAL ELEVATION MODEL
Elevation data is collected from the original NAPP photography using an instrument known as an analytical stereo plotter. Theoperator calibrates the instrument on the known control points in the photograph. The operator then locates points on the ground at approximately 90 meter intervals and establishes the elevation of that point. The data is then interpolated to a finer resolution and used in the ortho correction process. This results in approximately 4,000 points of elevation being collected for each DOQQ map.
The digital elevation data is provided to DNR in an ASCII file that can be used to automatically generate 10 foot topographic contour lines with an accuracy of + or - 5 feet.
ORTHO CORRECTION OF THE PHOTOGRAPH
Ortho correction is performed on an International Imaging Systems digital image processing system that is capable of integrating the scanned image with the aerotriangulation solution and digital elevation model. Maryland's DOQQ maps are produced in color which requires 3 separate bands that are individually corrected.
This process starts by scanning the NAPP photograph at 32 microns which is 800 dots per inch. During the ortho correction process, every pixel (4'x 4' on the ground) is moved to its correct position based on the systems ability to manipulate the image according to the aerotriangulation solution and digital elevation model. This process is similar to copying a photograph on a pliable compound and draping it over a wire-frame model of the topographic features while stretching or shrinking known points in the photograph to their correct locations.
The contractor delivers each separate band to MD DNR for future use in automated interpretation and classification of natural features. They also merge the three bands into a single color composite image that is used for all map processing at MD DNR. The separate bands and the color composite image each occupy 28 - 32 megabytes on the storage media. The file size was increased starting with the 1993 photography (Project Area 8) to provide complete photo coverage when converting from State Plane to UTM. The precise geographic coordinates of the map are transferred to the image files so that they are always properly registered.
MYLAR MAP PRODUCTION
DNR's contract requires that two 1:12,000 scale mylar maps be produced. One is given on a gratis basis to participating counties. The other is maintained by DNR for distribution of blueline maps. This product was specified so that inexpensive paper maps could be taken into the field for studies and annotations that would later be entered into GIS system applications.
The ortho corrected mylars were created using analog techniques until early 1994. Mylar products after this date are produced using digital methods. The original film products are maintained by the contractor for use in creating separate products. The plain ortho-corrected image with no annotations has been used for soil mapping activities by the SCS. These products will prove to be very valuable to agencies that have a need for ortho-corrected imagery but can not deal with digital products.
QUALITY ASSURANCE OF DOQQ MAPS
MD DNR performs quality control checks on the image quality and geo-referencing of each file. Image quality is checked by displaying the map image at large scales and scrolling around the image while looking for anomalies in the image. We also display the grid over the composite image and check the positional accuracy of the control points using the aerotriangulation report. If the color balance of the color composite image does not appear to be optimum, we make another composite that meets our production requirements.
MD DNR checks the mylar products for border block information, edge matching and image quality. When image quality is questioned, a blueline print of the mylar is made to determine the adequacy of the image. If mylars are unacceptable, they are returned to the contractor for production of new mylars.
In additon when the data is displayed at full resoultion, it is approximately 1:12000 scale, the mapped scale of the original one-meter product. The original files are available for distribution on hard medium such as CD-ROMs or DVD-ROMs. There is a charge involved for CD-ROMs or DVD-ROMs. Please contact MD DNR via the distributor information section or visit, <http://msgic.state.md.us/techtool> for further details.