Publisher Summary
This chapter discusses image processing techniques and algorithms in the physical context of remote sensing. The chapters also describes tools for spectral and spatial transforms that are useful in the correction and calibration of images for atmospheric and sensor effects. Spectral transformations alter the spectral space and spatial transformations alter the image space. Many of these transformed spaces are useful for thematic classification and are collectively called feature spaces in that context. A variety of spectral transformations are examined, ranging from nonlinear spectral band ratios to linear transformations of various types. Some are designed to improve quantitative analysis of remote-sensing images, while others simply enhance subtle information so that it is visible. The chapter concludes the following points: (1) spectral band ratios can help isolate spectral signatures from imagery by reducing topographic shading; (2) ratios involving near infrared (NIR) and red spectral bands are useful for vegetation measurements; (3) the principal components transform is optimal for data compression, but because it is data dependent, the resulting features have different interpretations for different images; (4) the tasseled-cap transform provides a fixed but sensor-specific transform that is based on soil and vegetation signatures; and (5) color–image composites can be enhanced for visual interpretation by a variety of spectral transforms, however, these transforms are not to be used for quantitative data analysis.
Research Article| February 01, 1990 Multispectral Ratio Selection Using Ternary Diagrams THOMAS A. KING; THOMAS A. KING Physical Science Laboratory, New Mexico State University, Las Cruces, NM 88003 Search for other works by this author on: GSW Google Scholar CHARLES E. GLASS; CHARLES E. GLASS Department of Mining and Geological Engineering, University of Arizona, Tucson, AZ 85721 Search for other works by this author on: GSW Google Scholar ROBERT A. SCHOWENGERDT ROBERT A. SCHOWENGERDT Department of Electrical and Computer Engineering, University of Arizona, Tucson, AZ 85721 Search for other works by this author on: GSW Google Scholar Environmental and Engineering Geoscience (1990) xxvii (1): 93–102. https://doi.org/10.2113/gseegeosci.xxvii.1.93 Article history first online: 09 Mar 2017 Cite View This Citation Add to Citation Manager Share Icon Share Twitter LinkedIn Tools Icon Tools Get Permissions Search Site Citation THOMAS A. KING, CHARLES E. GLASS, ROBERT A. SCHOWENGERDT; Multispectral Ratio Selection Using Ternary Diagrams. Environmental and Engineering Geoscience 1990;; xxvii (1): 93–102. doi: https://doi.org/10.2113/gseegeosci.xxvii.1.93 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu nav search search input Search input auto suggest search filter All ContentBy SocietyEnvironmental and Engineering Geoscience Search Advanced Search Abstract A method is presented to map spectral reflectance data of altered or unaltered rocks into color space to aid interpretation of color ratio composite (CRC) images for engineering geologic purposes. The method uses in situ spectral reflectance data or multispectral images of known lithologies, normalized ratio values, ternary diagrams, chromaticity diagrams, and color science theory to enable the engineering geologist to display spectral separability of lithologic units and improve the interpretation of CRC images. The normalized ratio values, derived from spectral reflectance data, can be plotted in a ternary diagram. Transforming normalized ratio values into equivalent chromaticity coordinates, the ternary diagram becomes a chromaticity diagram.From such a presentation, the engineering geologist can determine a priori the approximate color that each lithologic unit will appear in a CRC image. In this study, it was determined that the best CRC images correspond to the ternary/chromaticity diagrams that show the most separation of spectral reflectance data. This content is PDF only. Please click on the PDF icon to access. First Page Preview Close Modal You do not currently have access to this article.
Publisher Summary
This chapter focuses on spatial transforms, which can be combined with spectral transforms for certain applications such as image fusion and feature extraction for classification. Spatial transforms provide tools to extract or modify the spatial information in remote-sensing images. Some transforms such as convolution use only local image information, that is, within relatively small neighborhoods of a given pixel. Others, for example, the Fourier transform, use global spatial content. Between these two extremes, the increasingly important category of scale-space filters, including Gaussian and Laplacian pyramids and the wavelet transform provide data representations that allow access to spatial information over a wide range of scales, from local to global. The chapter concludes the following points: convolution and Fourier transform filtering are equivalent global processing techniques, except in the border region; a wide range of processing can be performed with small neighborhood windows, including noise removal and edge detection; scale-space filters allow access to image features according to their size, which is not possible with linear convolution or Fourier filters; and the resolution pyramid provides a unified description of scale-space filters, for example, Gaussian, Laplacian, and wavelet pyramids.
This chapter discusses common data models, providing the link among the physical remote sensing models, the sensor models, and image processing algorithms with examples. The connection among radiation, sensor models, and data models are also discussed and explored by simulation. The chapter concludes the following points: (1) the spectral statistics of remote-sensing image data are influenced by the topography in the scene and the topographic effect tends to correlate the data among spectral bands along a straight line through the origin of the reflectance scattergram; (2) the spectral statistics of remote-sensing image data are also influenced by the sensor's spectral passband locations and widths and noise characteristics; and random, spatially uncorrelated sensor noise increases the within-class variance of all surface materials equally; (3) the spatial and spectral statistics of remote-sensing image data are influenced by the sensor's spatial response function, which increases the spatial correlation length, reduces within-class variance, and creates mixed spectral vectors.
