« PreviousContinue »
By Alex Acosta
Color pictures generated from digital images are frequently used by geologists, foresters, range managers, and other professionals. These color products are preferred over black-and-white pictures because the human eye is more sensitive to color differences than to various shades of gray.
Color discrimination is a function of perception; because the colors in these color composites can be described subjec
tively, ambiguous color communication may result. Numerous color-coordinate systems are available that quantitatively relate amounts of red, green, and blue (RGB) in the form of digital triplets to the parameters of hue, saturation, and intensity perceived by the eye. Most of these systems implement a complex transformation of the RGB parameters to a color space that is hard to visualize, thus making it difficult to relate the triplets to perception parameters. Research at the USGS ADP Service Center in Flagstaff, Arizona, has developed a color-communication scheme that allows for accurate color communication and easy association of RGB triplets to the color generated by these triplets.
The color-communication scheme transforms RGB color space (fig. 3) into a chromaticity diagram (fig. 4). The diagram uses the (x,y) chromaticity diagram coordinates along with a third coordinate, z=1-x-y,
Figure 4. Chromaticity diagram with "hue" and "saturation." The chromaticity diagram is the area enclosed by the right triangle. The x chromaticity coordinate ranges from 0 to 1.0 on the horizontal side of the triangle. The y chromaticity coordinate ranges from 0 to 1.0 on the vertical side of the triangle. The hue is an angular value ranging clockwise from 0 to 360 degrees as indicated on the hue wheel. The hue wheel is centered at chromaticity coordinates (113,113). The saturation is a relative ratio measure along a hue spoke. Various values of saturation are given for hue = 240.
Figure 7. Examples of color enhancement of a Landsat Multispectral Scanner subscene (ID 2279-79367; near McMurdo Sound, Antarctica, obtained October 18, 1975. Band 4 = blue, band 5 = green, and band 7 = red. A, False-color composite. A 1-percent linear stretch (which linearly transforms the original data between the 1-percent and 99-percent cumulative frequency points into the entire 8-bit data range (0-255)) was applied to the data to enhance the color. B, Each band was stretched with digital number (DN) 0 mapped to DN 63 in an attempt to lighten the dark areas. C, Cube number (C) image file of A was stretched with DN 0 mapped to DN 63; the result was recombined with the original data. D, The (x,y,z) parameter image files of A were stretched with a 1-percent linear stretch; the result was combined with the C image file used in C and inversely transformed to RGB color space.
to generate a color triangle (fig. 5) and merges the color triangle, the chromaticity diagram, and a "hue" wheel, which is centered at chromaticity coordinates (1/3,1/3), to form a figure called a Chromaticity Color Triangle with Hue and Saturation (CCTHS) (fig. 6). This figure defines several parameters (x, y, "hue", and "saturation") that may be used in conjunction with other parameters (z and "cube number") to produce derivative digital images that can be used in various ways. For example, the (x,y,z) coordinates can be scaled in an 8-bit range to generate image files that produce enhanced color products (fig. 7). As another example, a fourth image file can
act as the associated "cube number" of an RGB color set to produce a color composite that depicts the merging of the fourth image with the color set (fig. 8).
Research in digital image color processing that has led to the development of the CCTHS has given scientists a tool that can be used for color-composite analysis, accurate color communication, and improved color processing. The CCTHS allows the user to relate graphically defined parameters to visually perceived colors. Because these parameters can be mathematically defined, they can be calculated and made available from RGB digital images for further processing.
Landsat Multispectral Images of Antarctica
By Baerbel K. Lucchitta and Jo-Ann Bowell
Scientists at the USGS Image Processing Facility in Flagstaff, Arizona, are conducting a program to provide enhanced multispectral scanner (MSS) Landsat images of Antarctica. From these images, the scientists will be able to furnish accurate planimetric, false-color compositeimage maps in polar stereographic projection. In turn, these image maps can be used to locate and delineate blue-ice areas for the collection of meteorites, to produce special-purpose maps showing selected features, and to provide synoptic views that aid in the detection and interpretation of glaciological features associated with ice sheets, outlet glaciers, ice streams, and ice shelves. Researchers will also be able to monitor changes in coastlines and glacial features, to correlate different types of digital cartographic data, and to provide structural information in areas of limited bedrock outcrop to aid in regional geologic interpretation.
About 170 false-color composite images of Antarctica, covering Victoria Land, the coastline of West Antarctica, the Antarctic Peninsula, and other selected areas have been produced. These images cover about 10% of the continent. Digitally enhanced color images composed of bands 4, 5, and 7 of Landsat MSS data are superior to the commonly used black-and-white images of MSS band 7. The color images show considerably more detail of the land surface. The images have undergone routine geometric correction and some noise removal. Because the data in many scenes are saturated in bands 4 and 5, new algorithms were devised to restore the saturated snow and ice information and to