Comparative Study and Analysis of Various Edge Detection Algorithms in Digital Image Processing
Journal Title: Scholars Journal of Engineering and Technology - Year 2013, Vol 1, Issue 2
Abstract
In the field of image processing, edge detection is an important step for extracting relevant and meaningful information from digital images. The main goal of edge detection techniques is to obtain and detect thin edges of the objects present in the image, so that the result is more suitable for further processing and analysis such as boundary detection, image segmentation, motion detection/estimation, texture analysis, object identification, feature detection, implementing various transformations and so on. We tested six edge detection algorithms that use different methods for detecting edges and compared their results under a variety of situations to determine a generally preferable technique under different sets of conditions. This data could then be used to create a multi-edge-detector system, which analyses the scene and runs the edge detector best suited for the current set of data. For each of these edge detectors we considered two different ways of implementation, the one using intensity only and the other coupling to it, the colour information. We also considered one additional edge detector which takes a different philosophy to edge detection. Rather than trying to find the ideal edge detector to apply to traditional photographs, it would be more efficient to merely change the method of photography to one which is more conducive to edge detection. It makes use of a camera that takes multiple images in rapid succession under different lighting conditions. It has been observed that the Canny’s edge detection algorithm performs better than all these operators under almost all scenarios. Evaluation of the images showed that under noisy conditions Canny, LoG( Laplacian of Gaussian), Robert, Prewitt, Sobel exhibit better performance, respectively. It has been observed that Canny’s edge detection algorithm is computationally more expensive compared to LoG( Laplacian of Gaussian), Sobel, Prewitt and Robert’s cross operator.
Authors and Affiliations
Sreemana Datta
Keywords
Related Articles
- EP ID EP384344
- DOI -
- Views 136
- Downloads 0
A New Non-monotone Self-Adaptive Trust Region Method with Fixed Step-size for Unconstrained Optimization
In this paper, we propose and analyze a new non-monotone self-adaptive trust region method with fixed step-size for unconstrained optimization. Unlike the traditional non-monotone trust region method, our algorithm utili...
The Analysis of Regulatory Strategy about Government-Invested Projects Based on the Imperfect Credibility Threats
In the construction market, it is very difficult to obtain the private information of project management units which have uneven aptitude for the government. In order to reduce the government’s risk caused by adverse sel...
The application research on the flipped classroom in the course of public art education curriculum based on the micro class
The flipped classroom is a new teaching method which has been widely praised in recent years, but it was rarely used in the public art education curriculum. The current situation, the nature, the function and the defects...
Environmental Monitoring and Greenhouse Control by Distributed Wireless Sensor Network
A sensor is a miniature component which gauge physical parameters from the environment. Sensors measure the physical parameters and transmits them either by wired or wireless means. In wireless medium the sensor and its...
Comparison of Moisture Removal Rate for Five Samples of Sliced Staple Food Using Multipurpose Convective Cabinet Dryer
This paper is aimed at studying the rate of moisture removal from samples of staple food using a multipurpose convective cabinet dryer. A multipurpose convective cabinet dryer is designed, fabricated and used to study th...