News |  Sitemap |  Contact
PDF Export  | 

Comparison of Edge Detection Operators

Abstract:
While the human brain can easily discern the edge of an object in an image it is viewing, a camera or other imaging system may have difficulty in identifying the same edge. The main problem with image analysis is the amount of information that must be processed. Therefore, analysis of digital images using software can often be simplified by applying various operators to reduce the amount of information. For this comparison, a single image was manipulated by applying the Sobel, Perwitt, Laplacian of Gaussian, Robinson, and Kirsh operators. The first step in the analysis was to create a histogram of the original image for comparison purposes. Next noise was added to the image to help differentiate the different operators. Two different noise options were used for this analysis: smoothing and salt and pepper. Then each operator was applied to the image with each of the noise options as well as without noise added. The results of this analysis are included in the provided poster presentation. As shown on the poster, each of the operators reduced the amount of information of the image, but they also had a tendency to reduce the edges of the objects as well. Additionally, adding noise helped in some instances but hindered in others. The best overall edge operator for this particular image was the Laplacian of Gaussian operator but to provide the optimal results smoothing noise must be added. When both are applied, the edges of each of the objects in the image are clearly discernable.

Keywords:
Comparing Edges

Author  
Michael Morrow

Organisation   
University of Alabama Birmingham

Homepage   

Short Biography   
I am Michael Morrow I have been writing in image j for one semester and am enrolled at the University of Alabama in Birmingham as a senior in my undergrad program. This Project was required by my professor to submit or final term projects to this image j conference. I chose computer science as my major 2 years ago when one of my friends introduced me to the material that was cover in the curriculum and I found it very interesting and decided that it was what I wanted to do the rest of my career. I also have a engineering background which is what I was working one prior to my Computer Science degree and am considering finishing that program up after a year of working in my field. I have recently been offered a job working for L-3 communications in Warner Robins Georgia and am extremely excited to continue my learning experience with their organization.

© Luxembourg Institute of Science and Technology | Legal Notice