Date of Award

12-1-1987

Document Type

Thesis

Degree Name

Master of Arts (MA)

Department

Geography and Geology

First Advisor

Dr. Michael P. Peterson

Abstract

The digital format of remote sensing data facilitates the measurement of spatial patterns. The concept of measurability of spatial patterns has important geographic implications and may open up an alternative applications for satellite remote sensing of urban areas. Texture analysis, a set of techniques developed in pattern recognition, is found to be useful in measuring spatial pattern on digital imagery. Two approaches of texture analysis are selected. One is Haralick's Spatial Dependence M atrix, the other is Jernigan's, et. al., Entropy-based texture measures. They perform in spatial domain and frequency domain respectively. Ten subimage areas in Omaha suburb are selected from a Landsat TM image. The subimage areas includes the major residential spatial patterns in the area, Through analysis, it is found that residential areas with d iffe re n t spatial features do present distinguishable texture measures, in both SPADEP and Entropy-based texture analysis. With, the introduction of texture analysis, a new set of terminology can be used to describe a spatial pattern and may greatly enhance our concepts of certain spatial phenomena. Potential application of texture analysis in this context could be in urban land use mapping, computer-assisted land use monitoring and comparative study in urban spatial patterns.

Comments

A Thesis Presented to the Department of Geography/Geology and the Faculty of the Graduate College University of Nebraska In Partial Fulfillment of the Requirements for the Degree Master of Arts University of Nebraska at Omaha. Copyright 1987, Li Bin.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

COinS