Supervised Constrained Optimization of Bayesian Nonlocal Means Filter With Sigma Preselection for Despeckling SAR Images

Gómez, Luis; Munteanu, Cristian; Buemi, María Elena; Jacobo-Berlles, Julio; Mejail, Marta

doi:10.1109/tgrs.2013.2269866

Cited by 17 publications

(9 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Speckle reduction (despeckling) is an important issue in SAR, and a plethora of excellent despeckling filters have been proposed based on local statistical and adaptive models since the very beginning [4][5][6]. More recently, new approaches include improvement of existing filters [7,8], Nonlocal Means methods [9][10][11], the Curvelet Transform [12,13], the Wavelet Transform [14][15][16], Partial Differential Equations formulation (such as Total Variation and Diffusion-Reaction models) [17,18], or combinations of methods employing Machine Learning techniques [19][20][21]. There are also methods devoted to the multilook case such as [22] which benefits from the sparsity characteristics of speckle to build a non-parametric statistical model for speckle reduction.…”

Section: Introductionmentioning

confidence: 99%

Statistical Properties of an Unassisted Image Quality Index for SAR Imagery

2019

Self Cite

View full text Add to dashboard Cite

The M estimator is a recently proposed image-quality index used to evaluate the despeckling operation in SAR (Synthetic Aperture Radar) data. It is used also to rank despeckling filters and to improve their design. As a difference with traditional image-quality estimators, it operates not on the filtered result but on a derived one, i.e., the ratio image. However, a deep statistical analysis of its properties remains open and, with it, the ability to use it as a test statistic. In this work, we focus on obtaining insights into its distribution as well as on exploring other remarkable statistical properties of this unassisted estimator. This study is performed through EDA (Exploratory Data Analysis) and the well-known ANOVA (ANalysis Of VAriance). We test our results on a set of simulated SAR data and provide guides to enrich the M estimator to extend its capabilities.

show abstract

Section: Introductionmentioning

confidence: 99%

Statistical Properties of an Unassisted Image Quality Index for SAR Imagery

2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…In author's experience, given some results from several despeckling filters, there are SAR experts that prefer one solution over the other [7]. Additionally, one may ask what the purpose of such filtering is, because, one solution may be desired in terms of the required image postprocessing operation (image classification and edge segmentation) or in terms of demanding a lower computational time.…”

Section: Some Results For Commonly Applied Despeckling Filtersmentioning

confidence: 99%

“…Sometimes, there is no special interest to obtain the optimal filtered image. In [7], the user guides the design of an NL-means Bayesian filter to get not an optimal design but his desired design, i.e., a subjective solution satisfying the user's criteria.…”

Section: Introductionmentioning

confidence: 99%

“…The proposed estimator operates on the ratio image obtained as the point-to-point ratio between the original image (noisy image) and the filtered image. The use of ratio images to evaluate a filtering operation is common in recent publications [5], [13], [7]. This is also becoming a standard analysis in the SAR community and it consists of visually inspecting the regularity pattern shown in the ratio image and estimating the ENL in a homogeneous area within this ratio image.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A New Image Quality Index for Objectively Evaluating Despeckling Filtering in SAR Images

Gómez

Buemi

Jacobo-Berlles

et al. 2016

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

Synthetic aperture radar (SAR) images are corrupted with a multiplicative granular-like noise pattern known as speckle. The goal for a despeckling filter consists of suppressing the speckle while preserving all the scene features such as texture, point-type targets, and, especially, edges. There exist several speckle filtering techniques and a relevant number of image quality indexes to evaluate the performances of a filtering operation on an SAR image. However, assessing the superiority of a filter over other is not a trivial issue. In this work, we present a new referenceless estimator (αβ-ratio estimator) based on the ratio edge detector which allows helping in objectively evaluating a filter realization on SAR images. The proposed estimator operates on the ratio image obtained as the point-to-point ratio between the original image (noisy image) and the filtered image. An ideal filter operation on an image implies that, in areas where speckle is fully developed, the ratio image should have the features of pure speckle and no geometric content. The new estimator measures the remaining geometric content within the ratio image. This new estimator is easy to compute and it provides an excellent metric to rank a filtering operation on real SAR images.

show abstract

“…Image segmentation is the division of an image into regions or categories, which correspond to different objects or parts of objects. A good segmentation is typically one in which pixels in the same category have similar grayscale of multivariate values and form a connected region or neighboring pixels which are in deferent categories have dissimilar values [6]. The goal of the segmentation process is to simplify the representation of an image into something which will be easier to analyze.…”

Section: Clustering Schemamentioning

confidence: 99%

SAR image denoising via clustering based Linear Discriminant Analysis

Rajapriyadharshini

2015

2015 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

View full text Add to dashboard Cite

De-speckling is an important problem in synthetic aperture radar (SAR) image analysis. Speckle noise in SAR images is assumed to be representation of a multiplicative noise. Digital images plays an significant role in daily life applications such as satellite television, images of magnetic resonance as well as in the area of study such as remote sensing applications, geographic information system. The data set which are collected by the image sensor are sometimes polluted or infected by noise. The Synthetic Aperture Radar (SAR) is very useful for providing information about earth's surface by using the relative motion between antenna and its target. However, these images are affected with granular noise termed as Speckle noise. Speckle removal is one of many cause to give efficient image visualization, thus remains a major issue in SAR image processing. To address this problem here proposed the clustering based schema along with Linear Discriminant Analysis(LDA). The denoising of SAR image is based on clustering the noisy image into several disjoint local regions which has similar spatial structure and then again denoise each splitted regions by using wiener filtering in domain Linear Discriminant Analysis(LDA).The experimental resultant shows that the denoised patches of all clusters were finally used to reconstruct the noise-free image.

show abstract

Supervised Constrained Optimization of Bayesian Nonlocal Means Filter With Sigma Preselection for Despeckling SAR Images

Cited by 17 publications

References 39 publications

Statistical Properties of an Unassisted Image Quality Index for SAR Imagery

Statistical Properties of an Unassisted Image Quality Index for SAR Imagery

A New Image Quality Index for Objectively Evaluating Despeckling Filtering in SAR Images

SAR image denoising via clustering based Linear Discriminant Analysis

Contact Info

Product

Resources

About