Speckle Reduction of PolSAR Images in Forest Regions Using Fast ICA Algorithm
- PDF / 2,159,141 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 74 Downloads / 191 Views
RESEARCH ARTICLE
Speckle Reduction of PolSAR Images in Forest Regions Using Fast ICA Algorithm Alireza Sharifi & Jalal Amini & Josaphat Tetuko Sri Sumantyo & Ryutaro Tateishi
Received: 16 January 2014 / Accepted: 8 October 2014 # Indian Society of Remote Sensing 2014
Abstract Advancement in Polarimetric synthetic aperture radar technology (PolSAR) and its ability to capture images in different polarizations, facilitate accessing large volumes of information from a scene. These images as well as conventional SAR images are degraded by speckle noise. Processing and reducing the speckle of all the images separately is time consuming. So applying methods to simultaneously process different channels information can be very helpful. Several methods have been proposed for speckle reduction and each has advantages and disadvantages. In this paper, a despeckling approach based on fast independent component analysis (Fast ICA) algorithm is proposed for improving of the results when polarimetric channels are added. The results show the more input images to Fast ICA algorithm leads to better separation between signal components and the speckle. Analysis of the results for the ALOS (Advanced Land Observing Satellite) PALSAR polarimetric images showed that combination of
A. Sharifi Department of Surveying and Geomatics Engineering, University of Tehran, Tehran, Iran 11155-4563 e-mail: [email protected] J. Amini (*) Department of Surveying and Geomatics Engineering, University of Tehran, Tehran, Iran 11155-4563 e-mail: [email protected] J. T. Sri Sumantyo Center for Environmental Remote Sensing, Chiba University, Chiba 263-8522, Japan e-mail: [email protected] R. Tateishi Center for Environmental Remote Sensing, Chiba University, Chiba 263-8522, Japan e-mail: [email protected]
polarimetric channels improved 37 % the ENL value in comparison with only using the HH (horizontal-horizontal)-, HV (horizontal-vertical)-, and VV (vertical-vertical)- polarized channels. Keywords Independent component analysis . PolSAR . Speckle . Remote sensing
Introduction Synthetic Aperture Radar (SAR) is an active remote sensing system typically mounted on a moving platform such as an aircraft or spacecraft. This system scores significantly compared to the other optical and infrared sensors because of some features such as ability to work in unfavorable weather conditions and in the night (Li et al. 2010). SAR polarimetry (PolSAR) has demonstrated, particularly during the last decade, its significance for the analysis and the characterization of the earth surface, as well as for the quantitative retrieval of biophysical and geophysical parameters. The capability to explore the complete space of polarization states represents one of the most important properties of PolSAR data, as optimization procedures may be foreseen. The second important property of PolSAR data is its inherent multidimensional nature that allows a more precise characterization of the scattering process at the resolution cell than single polarization da
Data Loading...
