USGS ScienceSearch

Geology topics

W.T. Borgeson

Publications and source records attributed to W.T. Borgeson.

3 recordsLinked to original sources

Geometric accuracy of Landsat-4 and Landsat-5 Thematic Mapper images.

The geometric accuracy of the Landsat Thematic Mappers was assessed by a linear least-square comparison of the positions of conspicuous ground features in digital images with their geographic locations as determined from 1:24 000-scale maps. For a Landsat-5 image, the single-dimension standard deviations of the standard digital product, and of this image with additional linear corrections, are 11.2 and 10.3 m, respectively (0.4 pixel). An F-test showed that skew and affine distortion corrections are not significant. At this level of accuracy, the granularity of the digital image and the probable inaccuracy of the 1:24 000 maps began to affect the precision of the comparison. The tested image, even with a moderate accuracy loss in the digital-to-graphic conversion, meets National Horizontal Map Accuracy standards for scales of 1:100 000 and smaller. Two Landsat-4 images, obtained with the Multispectral Scanner on and off, and processed by an interim software system, contain significant skew and affine distortions. -Authors

Photogrammetric Engineering and Remote Sensing

TESTS OF LOW-FREQUENCY GEOMETRIC DISTORTIONS IN LANDSAT 4 IMAGES.

Tests were performed to investigate the geometric characteristics of Landsat 4 images. The first set of tests was designed to determine the extent of image distortion caused by the physical process of writing the Landsat 4 images on film. The second was designed to characterize the geometric accuracies inherent in the digital images themselves. Test materials consisted of film images of test targets generated by the Laser Beam Recorders at Sioux Falls, the Optronics* Photowrite film writer at Goddard Space Flight Center, and digital image files of a strip 600 lines deep across the full width of band 5 of the Washington, D. C. Thematic Mapper scene. The tests were made by least-squares adjustment of an array of measured image points to a corresponding array of control points.

Conference Paper