A new approach for enhancing corner detection algorithm

Other Title(s)

طريقة جديدة لتحسين خوارزمية التقاط الزوايا

Dissertant

al-Sarayirah, Maryam Mujalli

Thesis advisor

al-Abbadi, Muhammad

University

Mutah University

Faculty

Information Technology College

Department

Computer Science Department

University Country

Jordan

Degree

Master

Degree Date

2017

English Abstract

The corner has much important information that contributes to the retrieval of shapes, and in general can be considered as the main points in the images; such points produce the convergence of the edges of the image with many angles and curvatures.

Therefore, corner is very important and it plays a large role in applications of object recognition, image matching, in addition to its effective role in the cutting of images and different capture systems.There are many techniques and algorithms which designed for this purpose, so designers seek to retrieve the corner at the lowest time in the high accuracy of recovery; some of these methods lack accuracies in retrieval, while others suffer from high complexity time.

Therefore, this study was made to design a high-resolution corner detection system, considering the time spent in retrieval.

The proposed approach optimized corner detection model (OCDM)aims to mark the largest possible number of true corners depending on four steps, 1st step deals with detecting continues edge lines of the image, the second step represents the detected edges as curvatures, then all corners are aggregated, and finally detect the last set of corners by eliminating of false corners and noise with minimum execution time.The study dealt with three main phases in the capture of the corner.

The first method was used to extract the edges and then the corner.

We have used five of the most famous methods for extracting the edges, which are (Sobel, Canny, Roberts, Log and Prewitt).

The second method was adopted to apply four filters to the image directly which are Sobel, unsharp, Log and Prewitt.

The third method adopted to extract the edges then applied the unsharp filter.

The reason for using of unsharp is due to its speed compared to other filtration methods.

The results showed the effectiveness of the proposed method compared to the other methods used in terms of precision and sensitivity, their values reached 100% when using Sobel and Prewitt edge detectors.

Main Subjects

Information Technology and Computer Science

No. of Pages

58

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-780525