Please use this identifier to cite or link to this item:
Efficient Materials Mapping Using Hyperspectral Imaging Data
|uhm_ms_3759_uh.pdf||Version for UH users||21.02 MB||Adobe PDF||View/Open|
|uhm_ms_3759_r.pdf||Version for non-UH users. Copying/Printing is not permitted||21.02 MB||Adobe PDF||View/Open|
|Title:||Efficient Materials Mapping Using Hyperspectral Imaging Data|
|Advisor:||Lucey, Paul G|
|Issue Date:||Dec 2002|
|Publisher:||University of Hawaii at Manoa|
|Abstract:||Hyperspectral images contain large amounts of spectral data. An efficient material identification (EMI) process can incorporate methods which reduce the amount of spectra analyzed in a hyperspectral image and interpret the image quickly while still maximizing the quality of interpretation of the image. The purposes of this study are to implement and evaluate an EMI process, determine ways to improve the process, and to implement and test those improvements. An EMI process using spectral endmember detection, linear unmixing, and automated spectral endmember material identification by spectral feature matching is used to analyze a visible near-infrared hyperspectral image of Kaneohe Bay, Hawaiʻi, a region containing a complex mixture of natural and manmade elements. The EMI technique is successfully applied. Evaluation of the resultant interpretation of the hyperspectral image reveals shortcomings in the EMI process in endmember detection and material identification. Particularly, some detected endmembers are spectral targets useful only for mapping a small portion of image, and the library material database of the feature matching algorithm is insufficiently matched to materials in the Kaneohe Bay scene. Two improvements to the spectral endmember detection technique used in the EMI process are proposed: target detection and masking and more than one evaluation of image pixels as potential spectral endmembers. These proposed improvements are incorporated into a subsequent analysis of the Kaneohe Bay scene, resulting in an improved material analysis of the scene. The improvement is primarily due to the incorporation of target detection only. The EMI process is also applied to a multispectral image of the Aristarchus Plateau on the Moon. Target masking is incorporated into endmember detection, and a different material identification algorithm, one based on radiative transfer theory, is used. Highly detailed maps of lunar mineralogy and rock types are produced which are consistent with previous spectral analyses of the Aristarchus region. These maps add to previous findings in detail and specificity of location and quantity of mineral and rock distributions on the Aristarchus Plateau.|
|Description:||xvi, 117 leaves|
|Rights:||All UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.|
|Appears in Collections:||M.S. - Geology and Geophysics|
Please contact email@example.com if you need this content in an alternative format.
Items in ScholarSpace are protected by copyright, with all rights reserved, unless otherwise indicated.