scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Speckle noise suppression in SAR images (Oil spill images) using wavelet based methods and ICA technique

TL;DR: In this article, the authors present speckle noise suppression techniques using wavelet decomposition and Independent Component Analysis (ICA) methods. And the performance of their methods is measured in terms of Peak Signal-to-Noise Ratio (PSNR) values.
Abstract: Synthetic Aperture Radar (SAR) images are inherently degraded due to the coherent nature of the scattering phenomena called speckle. The presence of speckle decreases the utility of the SAR images by reducing the ability to detect ground objects. It affects the quality of the image adversely and hampers the observation of vital and crucial information present in the image. In this paper, we present speckle noise suppression techniques using wavelet decomposition and ICA methods. The algorithms viz., Projection Onto Approximation Coefficients (POAC), Projection Onto Span Algorithm (POSA) and Independent Component Analysis (ICA) are implemented on real SAR images and their results are tabulated. A comparison is made with standard speckle filters such as Lee filter, Frost filters, etc. The performance of our methods is measured in terms of Peak Signal-to-Noise Ratio (PSNR) values.
Citations
More filters
Journal ArticleDOI
TL;DR: In this article, a viable method to implement oil spill detection and monitoring based on marine radar is proposed, where the adaptive median filter algorithm is improved to eliminate the radar shared-frequency interference by adding the identification of noise points and resetting the neighborhood window.
Abstract: A viable method to implement oil spill detection and monitoring based on marine radar is proposed. The primary data of this study are obtained from the X-band marine radar of the teaching–training ship, YUKUN, of the Dalian Maritime University on July 21, 2010, when a pipeline burst and an oil spill accident occurred at the Xingang Port in Dalian. Aiming at the working characteristics of marine radar, the adaptive median filter algorithm is improved to eliminate the radar shared-frequency interference by adding the identification of noise points and resetting the neighborhood window. A power attenuation correction method is proposed to solve the uneven distribution in resolution and echo intensity by acquiring the average power distribution of radar images simultaneously. Oil spill will be easily detected from different sea backgrounds after morphological processing, gray segmentation, and image smoothing. Comparison with the images extracted from a thermal infrared sensor on the same monitoring point demonstrates the validity of the extraction method for oil spill based on X-band marine radar.

20 citations

Journal ArticleDOI
TL;DR: Experimental investigation of the proposed DNN‐SNRT conducted based on TerraSAR‐X images confirmed the superior enhancement of image quality over comparable recent filters and proved that it is able to reduce noise and preserve edges during the image quality enhancement process.
Abstract: The speckle noise present in synthetic‐aperture radar (SAR) images is responsible for hindering the extraction of the exact information that needs to be utilized for potential remote sensing applications. Thus the quality of SAR images needs to be enhanced by removing speckle noise in an effective manner. In this paper, A Deep Neural Network‐based Speckle Noise Removal Technique (DNN‐SNRT) is proposed that utilizes the benefits of convolution and Long Short Term Memory‐based neural networks to enhance the quality of SAR images. The proposed DNN‐SNRT uses multiple radar intensity images that are archived from the specific area of interest to facilitate the self‐learning of the intensity features derived from the image patches. The proposed DNN‐SNRT incorporates a dual neural network to remove speckle noise and flexibly estimates the thresholds and weights to achieve an effective SAR image quality improvement. The proposed DNN‐SNRT is capable of automatically updating the intensity features of SAR images during the training process. Experimental investigation of the proposed DNN‐SNRT conducted based on TerraSAR‐X images confirmed the superior enhancement of image quality over comparable recent filters. The results of the DNN‐SNRT scheme were also proved that it is able to reduce noise and preserve edges during the image quality enhancement process.

14 citations


Cites methods from "Speckle noise suppression in SAR im..."

  • ...O(i) = (Gi1 ⊕ G i 2 ⊕ … ⊕ G i l+k−1) (4) The output of Convolution layer 1 that is derived in Equation (4) is then used as the input to Layer 2 of the LSTM layer....

    [...]

  • ...LS(MSE) = 1 n n∑ i=1 (Ti(t) − Mi(t))2 (13) The error for each patched input image is derived, based on Equation (14)....

    [...]

  • ...G = [g(1), g(2), … , gl+k−1] (3) The output O(i) ∈R(l− k +1)xq of this convolutional layer is concatenated with each of the ith dimensional feature map that are derived using Equation (4)....

    [...]

  • ...The utilized LSTM model in the proposed DNN-SNRT approach is comprised of multiple cells that, in turn, are comprised of individual forget gates that are used to and output gates that are used to derive the outputs based on Equations (5)–(10) respectively. i(t) = 𝜎(Wi[M(t−1),O(i)] + Ci) (5) g(t) = 𝜎(Wg[M(t−1),O(i)] + Cg) (6) O(t) = 𝜎(Wn[M(t−1),O(i)] + Co) (7) n(t) = tanh[Wh[M(t−1),O(i)] + Cf ) (8) D(t) = g(t)◦D(t−1) + i(t)◦n(t) (9) and M(t) = O(t)◦ tanh(D(t)) (10) Here, ‘𝜎’ is the sigmoid function that possesses gating values that range from 0 to 1 by (◦) highlighting the element pairwise multiplicative process....

    [...]

  • ...Then, the feature representation corresponding to each feature vector G∈Rl+ k −1 based on weight vector f ∈Rkxh is determined based on Equation (3)....

    [...]

Journal ArticleDOI
TL;DR: Cavrinceanu et al. as mentioned in this paper presented a quantitative and qualitative assessment of common synthetic aperture radar image despeckling methods, analyzing their performance when applied to images containing natural oil slicks.
Abstract: Synthetic Aperture Radar (SAR) is traditionally used in the identification, mapping, and analysis of petroleum slicks, regardless of their origin. On SAR images, oil slicks appear as dark patches that contrast with the brightness of the surrounding sea surface. This distinction allows for automated detection algorithms to be designed using computer vision methods for objective oil slick identification. Nevertheless, efficient interpretation of the SAR imagery by statistical analysis can be diminished due to the speckle effect present on SAR images, a granular artefact associated with the coherent nature of SAR, which visually degrades the image quality. In this study, a quantitative and qualitative assessment of common SAR image despeckling methods is presented, analyzing their performance when applied to images containing natural oil slicks. The assessment is performed on Copernicus Sentinel-1 images acquired with various temporal and environmental conditions. The assessment covers a diverse area of filters that employ Bayesian and non-linear statistics in the spatial, transform and wavelet domains, focusing on their demonstrated performance and capabilities for edge and texture retention. In summary, the results reveal that filters using local statistics in the spatial domain produce consistent desired effects. The novel SAR-BM3D algorithm can be used effectively, albeit with a higher computational demand. Supplementary material: Implementations of the speckle filters used in this paper are made available at: https://github.com/cavrinceanu/specklefilters under an MIT license. Image statistics data is available for Tables 3-11 at: https://doi.org/10.6084/m9.figshare.13010405 Thematic collection: This article is part of the Remote sensing for site investigations on Earth and other planets collection available at: https://www.lyellcollection.org/cc/remote-sensing-for-site-investigations-on-earth-and-other-planets
References
More filters
Journal ArticleDOI
TL;DR: The basic theory and applications of ICA are presented, and the goal is to find a linear representation of non-Gaussian data so that the components are statistically independent, or as independent as possible.

8,231 citations


"Speckle noise suppression in SAR im..." refers methods in this paper

  • ...Out of the best standard de-noising filters viz., Lee filter, Frost Filter, etc., POAC, POSA and ICA, a comparison is being stated in this work....

    [...]

  • ...The basic task in ICA is to estimate A and S using only the observed vector X....

    [...]

  • ...ICA From the outputs shown in Fig.5-34, it is noticed that, ICA output has good contrast and brightness features for oil spill, while POAC and POSA smoothens better while preserving edges and textures when compared with all the other methods....

    [...]

  • ...The linear equation which represents the ICA model can be written as X = A.S (12) where X is the observed vector of N samples (x1,x2,…….xn), A is called the mixing matrix and S is the statistically independent components(IC’s)....

    [...]

  • ...ICA The ICA [12] is an entrenched statistical signal processing technique which decomposes a set of multi-variate signals into a base of statistically independent data-vectors with the...

    [...]

Journal ArticleDOI
TL;DR: Using maximum entropy approximations of differential entropy, a family of new contrast (objective) functions for ICA enable both the estimation of the whole decomposition by minimizing mutual information, and estimation of individual independent components as projection pursuit directions.
Abstract: Independent component analysis (ICA) is a statistical method for transforming an observed multidimensional random vector into components that are statistically as independent from each other as possible. We use a combination of two different approaches for linear ICA: Comon's information theoretic approach and the projection pursuit approach. Using maximum entropy approximations of differential entropy, we introduce a family of new contrast functions for ICA. These contrast functions enable both the estimation of the whole decomposition by minimizing mutual information, and estimation of individual independent components as projection pursuit directions. The statistical properties of the estimators based on such contrast functions are analyzed under the assumption of the linear mixture model, and it is shown how to choose contrast functions that are robust and/or of minimum variance. Finally, we introduce simple fixed-point algorithms for practical optimization of the contrast functions.

6,144 citations

Journal ArticleDOI
TL;DR: A novel fast algorithm for independent component analysis is introduced, which can be used for blind source separation and feature extraction, and the convergence speed is shown to be cubic.
Abstract: We introduce a novel fast algorithm for independent component analysis, which can be used for blind source separation and feature extraction. We show how a neural network learning rule can be transformed into a fixedpoint iteration, which provides an algorithm that is very simple, does not depend on any user-defined parameters, and is fast to converge to the most accurate solution allowed by the data. The algorithm finds, one at a time, all nongaussian independent components, regardless of their probability distributions. The computations can be performed in either batch mode or a semiadaptive manner. The convergence of the algorithm is rigorously proved, and the convergence speed is shown to be cubic. Some comparisons to gradient-based algorithms are made, showing that the new algorithm is usually 10 to 100 times faster, sometimes giving the solution in just a few iterations.

3,215 citations

Book
02 Feb 2009
TL;DR: In this article, the authors used a two-dimensional time-frequency approach to evaluate the effect of speckle properties in SAR images and showed that the effect on the spatial correlation of the specckle sparseness of SAR images can be influenced by the number of multilook-processed SAR images.
Abstract: Overview of Polarimetric Radar Imaging Brief History of Polarimetric Radar Imaging SAR Image Formation: Summary Airborne and Space-Borne PolSAR Systems Description of the Remaining Chapters Electromagnetic Vector Wave and Polarization Descriptors Monochromatic Electromagnetic Plane Wave Polarization Ellipse Jones Vector Stokes Vector Wave Covariance Matrix Electromagnetic Vector Scattering Operators Polarimetric Back Scattering Sinclair S Matrix Scattering Target Vectors k and Omega Polarimetric Coherency T and Covariance C Matrices Polarimetric Mueller M and Kennaugh K Matrices Change of Polarimetric Basis Target Polarimetric Characterization PolSAR Speckle Statistics Fundamental Property of Speckle in SAR Images Speckle Statistics for Multilook-Processed SAR Images Texture Model and K Distribution Effect of Speckle Spatial Correlation Polarimetric and Interferometric SAR Speckle Statistics Phase Difference Distributions of Single-Look and Multilook PolSAR Data Multilook Product Distribution Joint Distribution of Multilook Si2 and Sj2 Multilook Intensity and Amplitude Ratio Distributions Verifications of Multilook PDFs K Distribution for Multilook Polarimetric Data Summary Appendices PolSAR Speckle Filtering Introduction to Speckle Filtering of SAR Imagery Filtering of Single Polarization SAR Data Review of Multipolarization Speckle Filtering Algorithms PolSAR Speckle Filtering Scattering Model-Based PolSAR Speckle Filter Introduction to the Polarimetric Target Decomposition Concept Introduction Dichotomy of the Kennaugh Matrix K Eigenvector-Based Decompositions Model-Based Decompositions Coherent Decompositions The H/A/a Polarimetric Decomposition Theorem Introduction Pure Target Case Probabilistic Model for Random Media Scattering Roll Invariance Property Polarimetric Scattering a Parameter Polarimetric Scattering Entropy (H) Polarimetric Scattering Anisotropy (A) Three-Dimensional H/A/a Classification Space New Eigenvalue-Based Parameters Speckle Filtering Effects on H/A/a PolSAR Terrain and Land-Use Classification Introduction Maximum Likelihood Classifier Based on Complex Gaussian Distribution Complex Wishart Classifier for Multilook PolSAR Data Characteristics of Wishart Distance Measure Supervised Classification Using Wishart Distance Measure Unsupervised Classification Based on Scattering Mechanisms and Wishart Classifier Scattering Model-Based Unsupervised Classification Quantitative Comparison of Classification Capability: Fully PolSAR versus Dual- and Single-Polarization SAR Pol-InSAR Forest Mapping and Classification Introduction Pol-InSAR Scattering Descriptors Forest Mapping and Forest Classification Appendix Selected PolSAR Applications Polarimetric Signature Analysis of Manmade Structures Polarization Orientation Angle Estimation and Applications Ocean Surface Remote Sensing with PolSAR Ionosphere Faraday Rotation Estimation PolSAR Interferometry for Forest Height Estimation Nonstationary Natural Media Analysis from PolSAR Data Using a Two-Dimensional Time-Frequency Approach Appendix A: Eigen Characteristics of Hermitian Matrix Appendix B: PolSARpro Software: The Polariemtric SAR Data Processing and Educational Toolbox Index

1,931 citations


"Speckle noise suppression in SAR im..." refers background in this paper

  • ...Basically, speckle noise has the nature of a multiplicative noise[3], which causes a pixel-to-pixel variation in intensities and this variation manifests itself as a granular noise pattern in SAR images [4, 5]....

    [...]

Journal ArticleDOI
01 Jan 1994
TL;DR: A critical evaluation of speckle suppression filters is made using a simulated SAR image as well as airborne and spaceborne SAR images and computational efficiency and implementation complexity are compared.
Abstract: Speckle, appearing in synthetic aperture radar (SAR) images as granular noise, is due to the interference of waves reflected from many elementary scatterers. Speckle in SAR images complicates the image interpretation problem by reducing the effectiveness of image segmentation and classification. To alleviate deleterious effects of speckle, various ways have been devised to suppress it. This paper surveys several better‐known speckle filtering algorithms. The concept of each filtering algorithm and the interrelationship between algorithms are discussed in detail. A set of performance criteria is established and comparisons are made for the effectiveness of these filters in speckle reduction and edge, line, and point target contrast preservation using a simulated SAR image as well as airborne and spaceborne SAR images. In addition, computational efficiency and implementation complexity are compared. This critical evaluation of speckle suppression filters is mostly new and is presented as a survey p...

570 citations