Graduate Thesis Or Dissertation
 

WalshStephenJoseph1978.pdf

Público Deposited

Conteúdo disponível para baixar

Baixar PDF
https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/dv13zx46p

Descriptions

Attribute NameValues
Creator
Abstract
  • The identification and mapping of surface cover types within Crater Lake National Park, Oregon, has been effectively completed through the utilization of LANDSAT digital data and NASA U-2 color infrared aerial photography. Classification of LANDSAT data for surface cover type identification and mapping was accomplished through use of the Interactive Digital Image Manipulation System (IDIMS) which provided enormous classification flexibility due to the interactive capability of IDIMS and its color video display. LANDSAT data extraction techniques employed in this research include the generation of grayscales; density slice; spectral reflectance plots versus pixel frequency and LANDSAT bands for low sun angle and medium to high sun angle LANDSAT digital tape dates; spectral relationship plots; and control clustering of training set selection. The use of these techniques produced training statistics which show "good" class discrimination for study area classification. A covariance matrix, weighted divergence, and distance between clusters were generated for each of the 59 training classes to determine if any of the classes were statistically related and should be combined. Slope angle, slope aspect, and surface cover type variation, and to a lesser degree, crown size and crown density are the main environmental factors which account for spectral reflectance variation of surface cover types within Crater Lake National Park. Through an understanding of the influence of environmental factors in the reflectance value of surface cover types, one can evaluate the quality of training statistics and the location of training areas in order to reduce mis-classification or non-classification possibilities. A regression analysis, analysis of variance, T-values, F-values, and beta values were used to determine the relative degree that certain environmental factors influence the spectral reflectance of surface cover types producing a change in the mean reflectance of surface cover types per LANDSAT band. LANDSAT classification accuracy levels for 12 major surface cover types within Crater Lake National Park range from 84 to 98 percent. Accuracy levels are inversely related to the level of detail sought and obtained. LANDSAT's primary advantages over NASA U-2 color infrared aerial photography for surface cover type identification and mapping include: repetitive coverage; computer compatibility of data without photographic digitizing; multi-spectral scanner capability; and the further extension of wavelengths into the near infrared. The NASA U-2 aerial photography's primary advantages (over LANDSAT) for surface cover type identification and mapping include: increased scale; greater resolution; and stereoscopic viewing. Stereoscopic viewing was accomplished through use of an Old Delft Scanning stereoscope which provided 1.5 to 4.5 magnification. "Ground truth" is the link between LANDSAT data, aerial photography, and actual ground conditions. "Ground truth" should measure and/or observe surface cover types and parameters which are capable of influencing the spectral reflectance as detected by LANDSAT. Geologic activity, soil conditions, vegetation, and climate have in the past and do in the present influence the spectral reflectance of surface cover types within Crater Lake National Park, Oregon.
Declaração de direitos

Relações

Parents:

Itens

Miniatura Título Data de carga Acesso Ações
file details WalshStephenJoseph1978.pdf 2017-08-17 Público Baixar