Document Type
Thesis - Open Access
Award Date
1976
Degree Name
Master of Science (MS)
Department / School
Electrical Engineering
Abstract
Detection of edges in imagery represented by the matrix of data and classification of the data surrounded by an edge are investigated. An edge detection algorithm is developed and used to locate edges in digitized imagery. An algorithm called the adjacency classifier is developed to classify all groups in adjacent data points that are surrounded by edges. These two algorithms incorporate three important characteristics of the gradient vector: high gradients are inherent at edges, low gradients are inherent at edges, low gradients are inherent within homogeneous objects and gradient vector directions are perpendicular to the edge direction. Computer programs which implement both algorithms are documented. Digitized images of LANDSAT-1 satellite multispectral scanner data are analyzed with both algorithms. The accuracy of the edge detection algorithm is evaluated for LANDSAT-1 satellite scenes of land-lake edges. The documentation for SYMOVER, a gray-tone mapping program developed to display digitized images or digital data with a line printer is included.
Library of Congress Subject Headings
Boundary value problems
Imaging systems
Vector analysis
Landsat satellites
Format
application/pdf
Number of Pages
120
Publisher
South Dakota State University
Recommended Citation
Russell, Michael John, "Edge Detection and Adjacency Classification in Digital Data Via the Gradient Vector" (1976). Electronic Theses and Dissertations. 4970.
https://openprairie.sdstate.edu/etd/4970