Synthetic Aperture Radar Image Denoising Research Articles

Multitemporal SAR image despeckling based on non-local theory

In this paper, a multitemporal SAR image despeckling based on non-local theory (NLG-MulSAR) algorithm is proposed, which is improved based on the basic framework of the ratio-based multitemporal SAR image denoising (RABASAR). The temporal and spatial information of a multitemporal SAR image is integrated. The super image and the ratio image acquisition part of the RABASAR algorithm are optimized by the NLG filtering algorithm. The NLG algorithm does not need to transform multiplicative noise into additive noise on a synthetic aperture radar (SAR) image and then filter it. The NLG algorithm uses the nonlinear method to eliminate the influence of strong noise points while preserving image edge features. Based on the number of pixels in the non-local image block, the NLG algorithm avoids the generation of fuzzy noise on the filtered image. In this study, we use seven Gaofen three SAR images captured at different times in the Beijing area as experimental data to evaluate the effect of filtering methods in terms of five objective parameters: signal-to-noise ratio, standard deviation, equivalent number of looks, radiative resolution, and speckle noise index. In addition, based on the ratio image, we propose an index, namely the filtering edge coefficient of a multitemporal SAR image, to evaluate the filtering edge retention characteristics of a multitemporal SAR image. The results show that compared with the RABASAR filtering algorithm, the proposed NLG-MulSAR filtering algorithm can better balance the relationship between multiplicative noise and texture detail information and attenuate the speckle while protecting the texture detail information on the SAR image.

Synthetic Aperture Radar Image Denoising Algorithm Based on Deep Learning

Synthetic Aperture Radar Image Denoising Algorithm Based on Deep Learning

SAR‐IRGF: A novel improved rolling guidance filter for the synthetic aperture radar image

Inspired by the framework of the rolling guidance filter (RGF), a novel improved RGF for synthetic aperture radar (SAR) images is proposed in this letter, named SAR-IRGF. Based on the model of additive noise, the RGF has the complete control of detailed smoothing under a scale measure for optical images. However, the speckle is well known as a multiplicative noise, which invalidates the RGF for SAR images. In this letter, an appropriate edge guidance is utilized to optimize the RGF for SAR image denoising. Compared with the original version, better convergence is achieved. Experimental results on natural Ku-band airborne SAR images show that the proposed SAR-IRGF reduces more speckle while sharpening the edges effectively. Simultaneously, the SAR-IRGF demonstrates a significant nearly 5.5-fold improvement than RGF in the equivalent number of looks (ENL).

A CNN-Based Self-Supervised Synthetic Aperture Radar Image Denoising Approach

Synthetic aperture radar (SAR) plays an essential role in earth observation and projection due to its capability to penetrate clouds, which makes it possible to monitor terrestrial surfaces under all weather conditions. Multiplicative noise often occurs in the SAR signal, hampering the retrieval of information from SAR imagery. Convolutional neural networks (CNNs) have been used in many computer vision tasks and are helpful in image denoising. However, current CNN-based denoising approaches inevitably lead to a ``washed out'' effect that loses spatial details. Another limitation is that most typical CNN-based denoising models require a noise-free image for training. To address these issues, we propose a novel end-to-end self-supervised SAR denoising model: Enhanced Noise2Noise (EN2N), which can be trained without a noise-free image. To enhance the quality of the result images, the perceptual features from a pre-learned CNN are introduced to restore the spatial details by a hybrid loss function. Experiments show that our proposed method outperforms the typical denoising methods in terms of noise reduction and feature preservation based on image quality metrics. Also, the new hybrid loss could enhance the spatial details significantly. The good performance maintains the robustness throughout time, which reduces the uncertainty in time-series SAR caused by random noise. Benefiting from optimization of graphics processing unit (GPU) and multi-threading, the proposed method has higher computation efficiency than traditional methods. This study demonstrates the great potential of using our self-supervised deep learning approaches for SAR image denoising in the future.

MRDDANet: A Multiscale Residual Dense Dual Attention Network for SAR Image Denoising

Synthetic aperture radar (SAR), due to its inherent characteristics, will produce speckle noise, which results in the deterioration of image quality, so the removal of speckle in SAR image is very important for the subsequent high-level image processing. In order to balance the relationship between denoising and texture preservation, we propose a multiscale residual dense dual attention network (MRDDANet) for SAR image denoising. This algorithm can effectively suppress the speckle while fully retaining the texture details of the image. In MRDDANet, shallow features are extracted from the noisy images by multiscale modules with different kernel sizes, and then, the extracted shallow features are mapped to the residual dense dual-attention network to obtain the deep features of SAR image. Finally, the final denoising image is generated through global residual learning. MRDDANet has advantages of both multiscale blocks and residual dense dual attention networks. The dense connection can fully extract features in the image, and the dual-channel attention enables MRDDANet to pay more attention to noise information, which is beneficial to remove noise and keep the details of the original image at the same time. Compared with state-of-the-art algorithms, the results of the experiment indicate that our method not only improves various objective indicators but also shows great advantages in visual effects.

A Siamese Template Matching Method for SAR and Optical Image

Template matching between synthetic aperture radar (SAR) and optical image is a meaningful and important topic for remote sensing tasks. In this letter, a deep learning framework based on Siamese structure and U-Net is proposed. First, a sparse 3-D transform-domain collaborative filtering (BM3D) for SAR image denoising is applied. Then, deep features of SAR and optical images are extracted from two branches of the proposed Siamese U-Net network. Finally, a cross correlation layer is used to generate the heatmap of image pairs based on the deep features extracted, and the peak maximal location of the heatmap is seen as the best matching location result. During training, a balanced loss between ground truth and prediction is used to get trained model weights in order to make a balance between positive samples and negative samples. The experiments on the same test dataset demonstrate that the proposed method improves the matching accuracy and precision to some degree in comparison with existing template matching methods for SAR and optical images.

Convolutional Neural Network and Guided Filtering for SAR Image Denoising

Coherent noise often interferes with synthetic aperture radar (SAR), which has a huge impact on subsequent processing and analysis. This paper puts forward a novel algorithm involving the convolutional neural network (CNN) and guided filtering for SAR image denoising, which combines the advantages of model-based optimization and discriminant learning and considers how to obtain the best image information and improve the resolution of the images. The advantages of proposed method are that, firstly, an SAR image is filtered via five different level denoisers to obtain five denoised images, in which the efficient and effective CNN denoiser prior is employed. Later, a guided filtering-based fusion algorithm is used to integrate the five denoised images into a final denoised image. The experimental results indicate that the algorithm cannot eliminate noise, but it does improve the visual effect of the image significantly, allowing it to outperform some recent denoising methods in this field.

SAR Image Denoising Algorithm Based on Bayes Wavelet Shrinkage and Fast Guided Filter

To remove the speckle noise of synthetic aperture radar (SAR) images, a novel denoising algorithm based on Bayes wavelet shrinkage and a fast guided filter is proposed. According to the statistical properties of SAR images, the noise-free signal and speckle noise in the wavelet domain are modeled as Laplace and Fisher-Tippett distributions respectively. Then a new wavelet shrinkage algorithm is obtained by adopting the Bayes maximum a posteriori estimation. Speckle noise in the high-frequency domain of SAR images is shrunk by this new wavelet shrinkage algorithm. As the wavelet coefficients of the low-frequency domain also contain some speckle noise, speckle noise in the low-frequency domain can be further filtered by the fast guided filter. The result of the denoising experiments of simulated SAR images and real SAR images demonstrate that the proposed algorithm has the ability to better denoise and preserve edge information.

SAR Image Denoising via Sparse Representation in Shearlet Domain Based on Continuous Cycle Spinning

How to suppress speckle noise effectively has become one of the key problems in remote sensing image processing. This problem also restricts the development of key technology severely, especially in military applications and so on. To overcome the shortcoming that the optimal solution of image denoising based on sparse representation does not have one-to-one mapping of the original signal space, in this paper, we propose a novel synthetic aperture radar (SAR) image denoising via sparse representation in Shearlet domain based on continuous cycle spinning. First, the Shearlet transform is applied to the noised SAR image. Second, a new optimal denoising model is constructed using the sparse representation model based on the cycle spinning theory. Finally, the alternate iteration algorithm is used to solve the optimal denoising model to obtain the denoised image. The experimental results show that the proposed method not only effectively suppresses the speckle noise and improves the peak signal-to-noise ratio of denoising SAR image, but also obviously improves the visual effect of the SAR image, especially by enhancing the texture of the SAR image.

SAR image de-noising based on texture strength and weighted nuclear norm minimization

As synthetic aperture radar (SAR) has been widely used nearly in every field, SAR image de-noising became a very important research field. A new SAR image de-noising method based on texture strength and weighted nuclear norm minimization (WNNM) is proposed. To implement blind de-noising, the accurate estimation of noise variance is very important. So far, it is still a challenge to estimate SAR image noise level accurately because of the rich texture. Principal component analysis (PCA) and the low rank patches selected by image texture strength are used to estimate the noise level. With the help of noise level, WNNM can be expected to SAR image de-noising. Experimental results show that the proposed method outperforms many excellent de-noising algorithms such as Bayes least squares-Gaussian scale mixtures (BLS-GSM) method, non-local means (NLM) filtering in terms of both quantitative measure and visual perception quality.

Scattering-Based Nonlocal Means SAR Despeckling

Speckle noise greatly limits both synthetic aperture radar (SAR) data human readability, especially for non-SAR-expert users, and performance of automatic processing and information retrieval procedures by computer programs. Therefore, despeckling of SAR images is an essential preprocessing step in SAR data analysis, processing, and modeling, as well as in information retrieval and inversion procedures. Up to now, one of the most accurate and promising despeckling approaches—among those based on a single SAR image—is the one relying on the nonlocal means concepts. However, at the best of our knowledge, most of the state of the art considers the despeckling problem only within a statistical framework, completely discarding the electromagnetic phenomena behind SAR imagery formation. In this paper, we introduce the novel idea of a physical-based despeckling, taking into account meaningful physical characteristics of the imaged scenes. This idea is realized via the implementation of a physical-oriented probabilistic patch-based (PPB) filter based on a priori knowledge of the underlying topography and analytical scattering models. This filter is suitable for SAR images of natural scenes presenting a significant topography. An adaptive version of the proposed scattering-based PPB filter for denoising of SAR images including both mountainous and flat areas is also developed. The performances of the proposed filter and its adaptive version are evaluated both qualitatively and quantitatively in numerical experiments using both simulated and actual SAR images. The proposed technique exhibits performance superior w.r.t. the standard PPB filter and comparable or, in some cases, superior to the state of the art, both in terms of speckle reduction and texture and detail preservation.

A New Approach for SAR Image Denoising

<p>In synthetic aperture radar (SAR) imaging, the transmitted pulses from space born antenna interacts with ground objects and returned energy or back scattered energy will be collected to get backscattered image. In this process, a speckle noise will be added because of the coherent imaging system and makes the study of SAR images very difficult. For better SAR image processing, the speckle has to be removed in the initial stages of processing and maintain all texture features efficiently. The BM3D method is generally considered as state of art method in denoising of SAR images. In this paper, it is proposed a technique to despeckle the speckle noise to the maximum extent while maintaining the edge characteristics.</p>

An Affinity-Based Algorithm in Nonsubsampled Contourlet Transform Domain: Application to Synthetic Aperture Radar Image Denoising

An algorithm in which affinity is merged into nonsubsampled contourlet transform (NSCT) to denoise synthetic aperture radar (SAR) image is proposed. The important information (boundaries or details) or non-important information (homogeneous area) in a SAR image is estimated based on affinity matrix, by which the affinity-based denoising assigns a high affinity element to the coefficients of NSCT that belong to same region. The foreground and background in NSCT domain are automatically initialized which avoids the need for user initialization. Foreground probability obtained by optimizing objective function can be used to achieve the posterior ratio. Combining the posterior ratio and prior ratio, we can obtain the shrunk coefficients. The proposed algorithm was applied to real SAR images denoising and compared through the SAR image vision effect, the equivalent number of looks (ENL) and the edge sustain index (ESI). Experimental results show that the proposed algorithm outperforms the compared algorithms and achieves the better denoising result and edge preservation.

Game Theory based Framework for Synthetic Aperture Radar Image De-noising and Change Detection

Game Theory based Framework for Synthetic Aperture Radar Image De-noising and Change Detection

A Modified BM3D Algorithm for SAR Image Despeckling

Abstract In synthetic aperture radar (SAR) imaging, pulses of microwave energy are transmitted towards the ground surface (target). The backscattered signal energy is measured at the receiving end. In this process, a speckle noise is added because of the coherent imaging system and makes the study of images very difficult. For better SAR image processing, the speckle has to be removed in the initial stages of processing and maintain all texture features efficiently. The (Block Matching 3D algorithm) BM3D method is generally considered as state of art method in denoising of SAR images. Still some amount of speckle left with images. This is due to sampling conversion at transmitter and receiver. In this paper, it is proposed a technique to despeckle the speckle noise to the maximum extent while maintaining the edge characteristics. This technique uses un-sampling at both the ends.

Study on Synthetic Aperture Radar Image Denoising Algorithm

In synthetic aperture radar (SAR) inherent speckle will affect the legibility of image details; give the image target detection adverse effects. In order to reduce the SAR image speckles noise, this article provided an improved algorithm based on median filter and wavelet semi-soft threshold shrinkage. First, reduced the SAR image speckle with median filtering method, then, with the filtered image, filtering the image with wavelet half soft threshold value contraction algorithm to noise, Simulation results show that the algorithm based on median filtering and improved algorithm of wavelet half soft threshold shrinkage of SAR image can better remove the speckle noise of the SAR image, while keep better edge, in Equivalent Number of Looks (ENL) and edge keep ability (FOM) aspects, it would be better than median filtering.

SAR Image Denoising Using an Improved Adaptive Bitateral Filter

Synthetic aperture radar (SAR) images are inherently affected by multiplicative speckle noise. The presence of speckle damages radiometric resolution, at the same time, it hampers the human interpretation and scene analysis for SAR images. On the base of studying and analyzing the mathematical model of the bilateral filter, the paper proposed a modified adaptive bilateral filter (MABF). First, it separates non-independent two-dimensional Gaussian filter into two independent one-dimensional Gaussian filter, which improves the operation speed greatly. Then through the effective noise parameter estimation, it adaptively selects optimal parameters, which improves the filtering effect. The real SAR image data is used to test the presented method and the experimental results verify that MABF is feasible and effective.

Wavelet Denoising of SAR Images Corrupted by Speckle Noise Using Cycle Spinning

A novel and efficient speckle noise reduction algorithm based on wavelet transform by cycle spinning for removing speckle of unknown variance and minimizing the effect of pseudo-Gibbs phenomena from Synthetic Aperture Radar (SAR) images is proposed. Therefore, we show that the sub-band decompositions of logarithmically transformed SAR images. Then, we process and reconstruct multi-resolution wavelet coefficients by wavelet-threshold using cycle spinning, a technique estimating the true images as the linear average of individual estimates derived from wavelet thresholded translated versions of the noise images. Experimental results show that the proposed de-noising algorithm is possible to achieve an excellent balance between suppresses speckle effectively and weaken as many image Gibbs phenomena as possible. Quantitative and qualitative comparisons of the results obtained by the new method with the results achieved from the other speckle noise reduction techniques demonstrate its higher performance for speckle reduction in SAR images.

SAR image denoising in nonsubsampled contourlet transform domain based on maximum a posteriori and non-local constraint

An approach of synthetic aperture radar (SAR) image denoising in nonsubsampled contourlet transform (NSCT) domain based on maximum a posteriori (MAP) and non-local (N-L) constraint is proposed. SAR image is firstly modelled by a nonlogarithmic additive model for modelling of the speckle in NSCT domain. Then, coefficients of real signals are obtained in the NSCT domain with MAP adaptive shrinkage. As it tends to eliminate too many coefficients that contain useful information by shrinkage, the N-L constraint is introduced to smooth the coefficients left in each subband, for each pixel in the subbands of NSCT corresponding to those in the same location of the original image. Experiments show that the proposed approach is effective in SAR image denoising and texture preserving, in comparison with some traditional algorithms.

A Nonlocal SAR Image Denoising Algorithm Based on LLMMSE Wavelet Shrinkage

We propose a novel despeckling algorithm for synthetic aperture radar (SAR) images based on the concepts of nonlocal filtering and wavelet-domain shrinkage. It follows the structure of the block-matching 3-D algorithm, recently proposed for additive white Gaussian noise denoising, but modifies its major processing steps in order to take into account the peculiarities of SAR images. A probabilistic similarity measure is used for the block-matching step, while the wavelet shrinkage is developed using an additive signal-dependent noise model and looking for the optimum local linear minimum-mean-square-error estimator in the wavelet domain. The proposed technique compares favorably w.r.t. several state-of-the-art reference techniques, with better results both in terms of signal-to-noise ratio (on simulated speckled images) and of perceived image quality.