Masters Theses

Date of Award

8-1989

Degree Type

Thesis

Degree Name

Master of Science

Major

Electrical Engineering

Major Professor

Dragana Brzakovic

Committee Members

M.A. Abidi, M.M. Trivedi

Abstract

This thesis investigates edge detection in gray-scale digital images by considering both the edge strength and the edge direction information in the neighborhood of each edge pixel in an image. The advantage of the proposed algorithm is that it is not sensitive to noise and it can maintain small edge details in a noisy environment.

The proposed edge detection algorithm is rule-based and utilizes multi template combinations. For each image pixel, a set of templates of different sizes is used to obtain the edge information. Then, the edge information, i.e.. the edge strengths and the edge directions, is processed using two rule-based criteria to produce the final edge strength and final edge direction of the image pixel. One criterion defines the magnitude relationships of qualified edge pixels. The other criterion defines the direction relationships of qualified edge pixels. Qualified edge pixels which pass both criteria constitute the edges of the given image. The performance of the proposed algorithm is examined using synthetic binary images and their various noisy versions and several infrared images.

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

Share

COinS