UMP Institutional Repository

Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image

Santoso, Ardhi Wicaksono (2017) Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image. Masters thesis, Universiti Malaysia Pahang.

[img]
Preview
PDF (Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image -Table of contents)
Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image -Table of contents.pdf - Accepted Version

Download (362kB) | Preview
[img]
Preview
PDF (Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image -Abstract)
Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image -Abstract.pdf - Accepted Version

Download (28kB) | Preview
[img]
Preview
PDF (Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image -References)
Fuzzy-based frost filter for speckle noise reduction of Synthetic Aperture Radar (SAR) image -References.pdf - Accepted Version

Download (174kB) | Preview

Abstract

The Synthetic Aperture Radar (SAR) image is a high-resolution image and is less influenced by weather conditions either day or night compared to the optical image. SAR image, because of its advantages, is becoming more popular than the optical image in the remote sensing area for earth observation study. However, the speckle noise that occurs in the SAR image causes difficulties in image interpretation, and speckle noise reduction process has become necessary before of the usage of SAR image. The ideal speckle filter has the capability of reducing speckle noise without losing the information and preserving its texture. This study proposes the use of speckle noise filter that as nearly possible to meet those criteria. This research has investigated the performance of existing filter, which was Frost, Lee, Kuan, and Median, and had applied it to ALOS-PALSAR images with homogeneous and heterogeneous earth area surfaces in Kuantan, Pahang, Malaysia. Filtered image is measured and evaluated using image quality parameters to show the performance of the filters in reducing speckle noise and preserving the texture. The parameter used for filters evaluation performances are Equivalent Number of Looks (ENL), Speckle Index (SI), Mean, Standard Deviation and Variance. The experiment results showed that Frost filter has better results compared to others and has been selected as the qualified existing filter. The Frost filter was modified by applying the fuzzy approach which was aimed at eliminating speckle noise while maintaining texture. There are four combinations of proposed filter, which are Frost-ATMAV, Frost-ATMED, Frost-TMAV, and Frost-TMED combination. Based on the results of comparison and evaluation of the filters, Frost-TMAV combination has been selected as the best-proposed filter. It had improved the performance of Frost filters for each parameter's measurement; it showed the improvement value of 19.47% for ENL, 8.48% for SI, 2.56% for mean, 6.15% for standard deviation and 2.00% for a variance, applied into homogeneous areas of ALOS-PALSAR images. While when used with heterogeneous areas, it improved 9.54% for ENL, 4.41% for SI, 3.03% for mean, 1.51% for standard deviation and 2.96% for the variance. It has been verified that the Frost-TMAV could be used for ALOS-PALSAR data pre-processing, which means that this filter can produce good-quality images based on parameters used when compared with other filters.

Item Type: Thesis (Masters)
Additional Information: Thesis (Master of Science) -- Universiti Malaysia Pahang – 2017; SV: DR. LUHUR BAYUAJI; NO CD: 10862
Uncontrolled Keywords: Synthetic Aperture Radar
Subjects: T Technology > T Technology (General)
Faculty/Division: Faculty of Computer System And Software Engineering
Depositing User: Ms. Nurezzatul Akmal Salleh
Date Deposited: 11 Jan 2018 03:00
Last Modified: 11 Jan 2018 03:00
URI: http://umpir.ump.edu.my/id/eprint/19681
Download Statistic: View Download Statistics

Actions (login required)

View Item View Item