|Home||Archived October 29, 2018||(i)|
Orthophotos combine the image characteristics of a photograph with the geometric qualities of a map. The primary digital orthophotoquadrangle (DOQ) is a 1-meter ground resolution, quarter-quadrangle (3.75 minutes of latitude by 3.75 minutes of longitude) image cast on the Universal Transverse Mercator projection (UTM) on the North American Datum of 1983 (NAD83). The geographic extent of the DOQ is equivalent to a quarter-quadrangle plus the overedge ranges from a minimum of 50 meters to a maximum of 300 meters beyond the extremes of the primary and secondary corner points. The overedge is included to facilitate tonal matching for mosaicking and for the placement of the NAD83 and secondary datum corner ticks. 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 the lines from north to south. The radiometric image brightness values are stored as 256 gray levels, ranging from 0 to 255.
The standard, uncompressed gray scale DOQ format contains an ASCII header followed by a series of 8-bit image data lines. The keyword-based, ASCII header may vary in the number of data entries. The header is affixed to the beginning of the image and is composed of strings of 80 characters with an asterisk (*) as character 79 and an invisible newline character as character 80. Each keyword string contains information for either identification, display, or registration of the image. Additional strings of blanks are added to the header so that the length of a header line equals the number of bytes in a line of image data. The header line will be equal in length to the length of an image line. If the sum of the byte count of the header is less than the sample count of one DOQ image line, then the remainder of the header is padded with the requisite number of 80 character blank entries, each terminated with an asterisk and newline character.
(1) Florida Department of Environmental Protection (FLDEP) (2) South Florida Water Management District (SFWMD) (3) U.S. Department of the Interior (DOI) (4) National Park Service (NPS)
Aerial cameras have current certification from the USGS, Geography Discipline, Optical Science Laboratory. Test calibration scans are performed on all source photograph scanners.
An in-house orthophoto accuracy software program is used to determine the horizontal positional accuracy for DOQ data produced by the Geography Discipline. The program determines the accuracy by finding the line and sample coordinates of the passpoints in the DOQ and fitting these to their ground coordinates to develop a Root Mean Square Error (RMSE) statistic. From 4 to 9 points are checked. As a further accuracy test, the image line and sample coordinates of the DEM corners are transformed and compared with the actual X,Y DEM corner values to determine if they are within the RMSE. Additional information on this testing procedure can be found in U.S. Department of the Interior, U.S. Geological Survey, 1993, Technical Instructions, ORACC Users Manual (draft): Reston, VA. DOQ's produced by cooperators and contractors use similarly approved RMSE test procedures.
Adjacent DOQ's, when displayed together in a common planimetric coordinate system, may exhibit positional discrepancies across common DOQ boundaries. Linear features, such as streets, may not be continuous. These edge mismatches, however, still conform to positional horizontal accuracy within the National Map Accuracy standard (NMAS).
The production procedures, instrumentation, and hardware and software used in the collection of color-infrared USGS DOQ's vary depending on systems used at the contract, cooperator, or USGS production sites. The description below details, in general, the process used in the production of the USGS Color Infrared (CIR) DOQ data sets collected as part of the South Florida ecosystem research effort.
The rectification process required, as input, a user parameter file to control the rectification process, a digital elevation model (DEM) gridded to user-specified bounds, projection, zone, datum and X-Y units, a scanned digital image file (PHOTO) covering the same area as the DEM, ground X-Y-Z point values (CONTROL_INPUT) and their conjugate photograph coordinates in the camera coordinate system, and measurements of the fiducial marks (CAMERA_INPUT) in the digitized image.
The camera calibration report (CAMERA_INPUT) provided the focal length of the camera and the distances in millimeters from the camera's optical center to the camera's eight fiducial marks. These marks defined the frame of reference for spatial measurements made from the photograph. Ground control points (CONTROL_INPUT) acquired from ground surveys or developed in aerotriangulation are third order class 1 or better and meet National Map Accuracy Standard (NMAS) for 1:12,000-scale. Ground control points were in the Universal Transverse Mercator on NAD83. Horizontal and vertical residuals of aerotriangulated tie-points were equal to or less than 2.5 meters. Standard aerotriangulation passpoint configuration consisted of nine ground control points, one near each corner, one at the center near each side, and one near the center of the photograph. The conjugate positions of the ground control points on the photograph were measured and recorded in camera coordinates.
The raster image file (PHOTO) was created by scanning an aerial photograph film diapositive with a precision image scanner. An aperture of approximately 25 to 32 microns was used, with an aperture no greater than 32 microns permitted. Using 1:40,000-scale photographs, a 25-micron scan aperture equated to a ground resolution of 1-meter. The scanner converted the photographic image densities to gray scale values ranging from 0 to 255 for each color band. Scan files with ground resolution less than 1 meter or greater than 1 meter but less than 1.28 meters were resampled to 1 meter.
The principal elevation data source (DEM) was a standard DEM data set from the National Digital Cartographic Data Base (NDCDB). The DEM used in the production of DOQ's generally has a 30-meter grid post spacing and possesses a vertical RMSE of 7 meters or less. A DEM covering the extent of the photograph was used for the rectification. The DEM was traversed from user-selected minimum to maximum X-Y values, and the DEM X-Y-Z values are used to find pixel coordinates in the digitized photograph using the transformations mentioned above. For each raster image cell subdivision, a brightness or gray-scale value was obtained using nearest neighbor, bilinear, or cubic convolution resampling of the scanned image. The pixel processing algorithm is indicated in the header file. An inverse transformation relateed the image coordinates referenced to the fiducial coordinate space back to scanner coordinate space. For those areas for which a 7.5-minute DEM is unavailable and relief differences are less than 150 feet, a planar-DEM (slope-plane substitute grid) may have been used.
Rectification Process: The photo control points and focal length were iteratively fitted to their conjugate ground control points using a single photo space resection equation. The camera location and orientation in the form of a rotation matrix was obtained from this mathematical fit. This rotation matrix was then used to find the photograph or camera coordinates of any other ground X-Y-Z point. Next, a two-dimensional fit was made between the measured fiducial marks on the digitized photograph and their conjugate camera coordinates. Transformation constants were developed from the fit, and the camera or photograph coordinates were used in reverse to find their conjugate pixel coordinates on the digitized photograph.
The DEM was traversed from user-selected minimum to maximum X-Y values, and the DEM X-Y-Z values were used to find pixel coordinates in the digitized photograph using the transformations mentioned above. For each raster image pixel, a brightness or gray-scale value was obtained using nearest neighbor, bilinear, or cubic convolution resampling of the scanned image. An inverse transformation related the image coordinates referenced to the fiducial coordinate space back to scanner coordinate space.
Quality Control: All data were inspected according to a quality control plan. During the initial production phase, all rectification inputs and DOQ data sets were inspected for conformance to standards. All DOQ's are tested for physical format standards.
The uncompressed DOQ is a raw binary file preceded by a metadata header that consists of keyword entries and blank entries to equal the length of a single or a multiple of a single line of image data. Each keyword entry in the header is 80 characters in length and terminated by an ASCII newline character included in the character count of the line.
|Home||Archived October 29, 2018|