Cellular automation for efficient impulse noise removal and edge detection using graphic processor unit
Low cost sensors allow integration of image and video features in consumer devices. Increasing image resolution and pixel dimension requires high performance image processing technique preferably one that could be parallelize. In the last few years, Graphics Processing Units have evolved into more f...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2010
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/40949/1/FK%202010%2067R.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Low cost sensors allow integration of image and video features in consumer devices. Increasing image resolution and pixel dimension requires high performance image processing technique preferably one that could be parallelize. In the last few years, Graphics Processing Units have evolved into more flexible and
powerful data-parallel processors. Graphics Processing Units are economical and are advantageous in a wide variety of computer architecture. Recent developments in programmability and rapid growth in the performance of graphic hardware, has provided groundwork for using the architecture in alternative domains of graphics applications. In this respect, Cellular Automata could handle this requirement with its parallel architecture. The characteristic of cellular automata is found to be highly suitable for vector processor such as the Graphic Processor Unit and Field Programmable Gate Array. In recent years the tendency to use of cellular automata in solving the problems of image processing has increased. Noise removal and edge detection are fundamental operations, which commonly applied as pre-processing step before subsequent image processing tasks. One of the significant factors which degrade the performance of edge detector method is This thesis present Cellular Automata models for noise removal and edge detection of the distorted image by salt and pepper noise. In order to enhance the performance of the Cellular Automata model, a Graphic Processor Units programming approach has been adopted. The results obtained show that the implemented Cellular Automata models are able to suppress noise and edge extraction in high noise intensity to 90 percents. The Cellular Automata models implemented on Graphic Processing Units have successfully outperformed the method implemented on Central processing Unit by factor of 2 for gray scale image and factor of 10 for colour images. The results indicate that cellular automata executed on the Graphic Processor Units build a solid foundation for the wide variety of application in image processing. |
|---|