Edge detection is a research field within Image processing and Computer vision, in particular within the area of feature extraction. It is extensively used in image segmentation when we want to divide the image into areas corresponding to different objects. Representing an image by its edges has the further advantage that the amount of data is reduced significantly while retaining most of the image information.
Since edges consist of mainly high frequencies, we can, in theory, detect edges by applying a high pass frequency filter in the Fourier domain or by convolving the image with an appropriate kernel in the spatial domain. In practice, edge detection is performed in the spatial domain, because it is computationally less expensive and often yields better results. Since edges correspond to strong illumination gradients, we can highlight them by calculating the derivatives of the image.
The present Thesis aims at extracting a good & accurate edge detected image from the application of various masks or edge detection operators on the image.Convolution is the mathematical tool, that is used to implement the various masks operators to get an edge detected image from the original image. Our thesis provides the implementation of the following edge detection techniques to get a better edge detected image: a) 1 Dimensional operators : Kirch ,Prewitt, Sobel and Quick Masking; b) 2 Dimensional operators: LOG( Laplacian of Guassian) and DOG( difference of Guassian).
Table of Contents
1. Introduction
2. Review of technology & Literature Survey
2.1 Edge detection
2.2 Images
2.2.1 Binary Images
2.2.2 Color Images
2.2.2(a) 8-Bit Color Images
2.2.2(b) 24-bit Color Images
2.2.2(c) True Color Images
2.2.2(d) BMP computer Image
2.3 Convolution
2.4 Edge Detectors
2.4.1 Sobel
2.4.2 Compass edge Detector/Prewitt
2.4.3 Laplacian of Gaussian(LOG)
2.4.4 Difference of Gaussian(DOG)
2.4.5 Laplacian for Edge Detection
3. Design & Architecture
3.1 Process Flow
3.2 System Architecture
3.3 Design of Architectural Module
4. Implementation
4.1 Implementation Environment
4.1.1 Introduction to JMF
4.1.2 Supported Content types of JMF
4.1.3 Component Architecture
4.2 Graphic User Interface(GUI)
5. Results and Analysis
6. Conclusion & Future Enhancements
Research Objectives and Key Topics
The primary objective of this thesis is to implement and evaluate the performance of various edge detection algorithms on digital images. By utilizing mathematical convolution tools and specific masks or operators, the research aims to extract accurate edge data, thereby reducing overall image data while preserving critical structural information.
- Comparison of 1D and 2D edge detection operators.
- Implementation of Kirsch, Prewitt, Sobel, and Quick Masking algorithms.
- Application of advanced 2D operators: Laplacian of Gaussian (LoG) and Difference of Gaussian (DoG).
- Performance evaluation through subjective human judgment and objective analysis (RMSE, SNR).
- Development of a Java-based GUI for interactive edge detection processing.
Auszug aus dem Buch
CHAPTER 1 INTRODUCTION
Edge detection is one of the fundamental operations in image processing. Many other operations are based on edge detection and much has been written about this subject. Detecting edges is a basic operation in image processing. The edges of items in an image hold much of the information in the image. The edges tell you where items are, their size, shape, and something about their texture. An edge is where the gray level of the image moves from an area of low values to high values or vice versa. The edge itself is at the center of this transition. An edge detector should return an image with gray levels. The detected edge gives a bright spot at the edge and dark areas everywhere else. This means it is the slope or rate of change of the gray levels in the edge. The slope of the edge is always positive or zero, and it reaches its maximum at the edge. For this reason, edge detection is often called image differentiation The problem in edge detection is how to calculate the derivative (the slope) of an image in all directions? Convolution of the image with masks is the most often used technique of doing this.
Summary of Chapters
1. Introduction: Introduces edge detection as a fundamental image processing operation and explains the basic concept of using convolution masks for differentiation.
2. Review of technology & Literature Survey: Covers the theoretical background of edge detection, intensity gradients, and the distinction between search-based and zero-crossing methods.
3. Design & Architecture: Outlines the system flow, from image acquisition and preprocessing to the modular application of various edge detection algorithms.
4. Implementation: Details the practical development using Java and the Java Media Framework (JMF), including the creation of the graphical user interface.
5. Results and Analysis: Presents the comparative evaluation of algorithms using both subjective assessment and objective metrics like SNR and RMSE.
6. Conclusion & Future Enhancements: Summarizes the findings regarding algorithm robustness against noise and suggests potential future improvements like multiscale detection.
Keywords
Edge Detection, Image Processing, Convolution, Feature Extraction, Sobel, Prewitt, Kirsch, Laplacian of Gaussian, Difference of Gaussian, JMF, Java, Grayscale, SNR, RMSE, Subjective Analysis
Frequently Asked Questions
What is the core focus of this research?
The research focuses on the implementation and performance study of various edge detection algorithms in digital image processing, exploring their effectiveness in different scenarios.
Which edge detection techniques are analyzed?
The study includes 1D operators (Kirsch, Prewitt, Sobel, Quick Masking) and 2D operators (Laplacian of Gaussian - LoG, and Difference of Gaussian - DoG).
What is the primary objective of this work?
The primary objective is to extract accurate edge data from images while evaluating which operators perform best under varying conditions such as noise, blur, and image density.
What methodology is employed for evaluation?
The thesis utilizes both subjective analysis (human evaluation) and objective metrics, including Signal-to-Noise Ratio (SNR) and Root Mean Square Error (RMSE), to assess algorithm performance.
What is covered in the main body of the text?
The main body details the theoretical background of edge detection, the architectural design of the software, the Java-based implementation, and a comprehensive analysis of the experimental results.
Which keywords best describe this study?
Key terms include Edge Detection, Convolution, Image Processing, Sobel, Prewitt, Laplacian of Gaussian, and performance evaluation.
Why are LoG and DoG considered 2D operators?
These operators are considered 2D because they utilize a 2D convolution kernel that combines Gaussian smoothing with Laplacian differentiation to capture edges, effectively handling noise better than simpler gradient methods.
How does the GUI contribute to this project?
The GUI provides an interactive environment that allows users to select input images and easily apply different edge detection algorithms, facilitating direct visual comparison of the output.
- Citar trabajo
- Ashima Kalra (Autor), 2008, Implementation and Performance Study of Edge Detection of Images, Múnich, GRIN Verlag, https://www.grin.com/document/496900
