The Development and Implementation of Classical Edge Detection Algorithms

No Thumbnail Available
Authors
Mancini, Ethan
Advisor
Cline, Kelly
Editor
Date of Issue
2023-04-28
Subject Keywords
Publisher
Citation
Series/Report No.
item.page.identifier
Title
The Development and Implementation of Classical Edge Detection Algorithms
Other Titles
Type
Presentation
Description
Abstract
Within the field of computer vision, edge detection was developed in the 1960s as a method of finding borders within images by using Taylor series to approximate the gradient of a grayscale image. Within the context of mathematics, we review the development of these early edge detection methods including the Canny, Sobel, and Prewitt operators. We implement these algorithms with computer code for demonstration purposes and compare the effectiveness of these early methods on modern applications of edge detection such as quality control in manufacturing and tree-ring counting. Additionally, we consider the development of historical methods for improving the usefulness of outputs including automated edge removal through thresholding using Ostu’s Method and pre-filtering using a Gaussian filter. We evaluate these historical methods after implementation in code on several edge detection operators.
Sponsors
Degree Awarded
Semester
Spring
Department
Mathematics