Pattern Strengthened Deep Model for Sar Image Classification

Convolutional neural networks (CNNs) have been applied to learn spatial features for high-resolution (HR) synthetic aperture radar (SAR) image classification. However, there has been little work on integrating the unique statistical distributions of SAR images which can reveal physical properties of terrain objects, into CNNs in a supervised feature learning framework. To address this problem, a novel end-to-end supervised classification method is proposed for HR SAR images by considering both spatial context and statistical features. First, to extract more effective spatial features from SAR images, a new deep spatial context encoder network (DSCEN) is proposed, which is a lightweight structure and can be effectively trained with a small number of samples. Meanwhile, to enhance the diversity of statistics, the nonstationary joint statistical model (NS-JSM) is adopted to form the global statistical features. Specifically, SAR images are transformed into the Gabor wavelet domain and the produced multi-subbands magnitudes and phases are modeled by the log-normal and uniform distribution. The covariance matrix is further utilized to capture the inter-scale and intra-scale nonstationary correlation between the statistical subbands and make the joint statistical features more compact and distinguishable. Considering complementary advantages, a feature fusion network (Fusion-Net) base on group compression and smooth normalization is constructed to embed the statistical features into the spatial features and optimize the fusion feature representation. As a result, our model can learn the discriminative features and improve the final classification performance. Experiments on four HR SAR images validate the superiority of the proposed method over other related algorithms.

The convolutional neural network (CNN) has shown great potential in many fields; however, transferring this potential to synthetic aperture radar (SAR) image interpretation is still a challenging task. The coherent imaging mechanism causes the SAR signal to present strong fluctuations, and this randomness property calls for many degrees of freedom (DoFs) for the SAR image description. In this paper, a statistics learning network (SLN) based on the quadratic form is presented. The statistical features are expected to be fitted in the SLN for SAR image representation. (i) Relying on the quadratic form in linear algebra theory, a quadratic primitive is developed to comprehensively learn the elementary statistical features. This primitive is an extension to the convolutional primitive that involves both nonlinear and linear transformations and provides more flexibility in feature extraction. (ii) With the aid of this quadratic primitive, the SLN is proposed for the classification task. In the SLN, different types of statistics of SAR images are automatically extracted for representation. Experimental results on three datasets show that the SLN outperforms a standard CNN and traditional texture-based methods and has potential for SAR image classification.

Thanks to the availability of large-scale data, deep Convolutional Neural Networks (CNNs) have witnessed success in various applications of computer vision. However, the performance of CNNs on Synthetic Aperture Radar (SAR) image classification is unsatisfactory due to the lack of well-labeled SAR data, as well as the differences in imaging mechanisms between SAR images and optical images. Therefore, this paper addresses the problem of SAR image classification by employing the Generative Adversarial Network (GAN) to produce more labeled SAR data. We propose special GANs for generating SAR images to be used in the training process. First, we incorporate the quadratic operation into the GAN, extending the convolution to make the discriminator better represent the SAR data; second, the statistical characteristics of SAR images are integrated into the GAN to make its value function more reasonable; finally, two types of parallel connected GANs are designed, one of which we call PWGAN, combining the Deep Convolutional GAN (DCGAN) and Wasserstein GAN with Gradient Penalty (WGAN-GP) together in the structure, and the other, which we call CNN-PGAN, applying a pre-trained CNN as a discriminator to the parallel GAN. Both PWGAN and CNN-PGAN consist of a number of discriminators and generators according to the number of target categories. Experimental results on the TerraSAR-X single polarization dataset demonstrate the effectiveness of the proposed method.

ABSTRACTAutomatic Target Recognition (ATR) based on Synthetic Aperture Radar (SAR) images plays a key role in military applications. However, there are difficulties with this traditional recognition method. Principally, it is a challenge to design robust features and classifiers for different SAR images. Although Convolutional Neural Networks (CNNs) are very successful in many image classification tasks, building a deep network with limited labeled data remains a problem. The topologies of CNNs like the fully connected structure will lead to redundant parameters and the negligence of channel-wise information flow. A novel CNNs approach, called Group Squeeze Excitation Sparsely Connected Convolutional Networks (GSESCNNs), is therefore proposed as a solution. The group squeeze excitation performs dynamic channel-wise feature recalibration with less parameters than squeeze excitation. Sparsely connected convolutional networks are a more efficient way to operate the concatenation of feature maps from different layers. Experimental results on Moving and Stationary Target Acquisition and Recognition (MSTAR) SAR images, demonstrate that this approach achieves, at 99.79%, the best prediction accuracy, outperforming the most common skip connection models, such as Residual Networks and Densely Connected Convolutional Networks, as well as other methods reported in the MSTAR dataset.

The classification of high-resolution (HR) synthetic aperture radar (SAR) images is of great importance for SAR scene interpretation and application. However, the presence of intricate spatial structural patterns and complex statistical nature makes SAR image classification a challenging task, especially in the case of limited labeled SAR data. This paper proposes a novel HR SAR image classification method, using a multi-scale deep feature fusion network and covariance pooling manifold network (MFFN-CPMN). MFFN-CPMN combines the advantages of local spatial features and global statistical properties and considers the multi-feature information fusion of SAR images in representation learning. First, we propose a Gabor-filtering-based multi-scale feature fusion network (MFFN) to capture the spatial pattern and get the discriminative features of SAR images. The MFFN belongs to a deep convolutional neural network (CNN). To make full use of a large amount of unlabeled data, the weights of each layer of MFFN are optimized by unsupervised denoising dual-sparse encoder. Moreover, the feature fusion strategy in MFFN can effectively exploit the complementary information between different levels and different scales. Second, we utilize a covariance pooling manifold network to extract further the global second-order statistics of SAR images over the fusional feature maps. Finally, the obtained covariance descriptor is more distinct for various land covers. Experimental results on four HR SAR images demonstrate the effectiveness of the proposed method and achieve promising results over other related algorithms.

Compressed sensing (CS) is currently utilized in the application of Synthetic Aperture Radar (SAR) imaging. Unfortunately, the CS SAR imaging performance degrades as SNR decreases. However, there exists a denoising convolutional neural network (DnCNN) based on 2D convolution (DnCNN-2D) in the field of image denoising. Traditional DnCNN-2D is not directly applicable to SAR imaging. Thus, a DnCNN based on 1D convolution (DnCNN-1D) for signal denoising is proposed and employed to denoise the SAR signal. The denoised SAR signal is subsequently utilized for CS SAR imaging. Each convolutional layer of DnCNN-1D has a fixed number and fixed size of filters and activation functions. The residual learning strategy and batch normalization are adopted in the DnCNN-1D. The segmental signal-to-noise ratio (SEGSNR) and the normalized mean square error (NMSE) are selected to evaluate signal denoising and SAR imaging performance, respectively. The experimental results indicate that imaging performance is significantly improved after denoising the SAR signal by the proposed method.

Synthetic Aperture Radar (SAR) images are prone to be contaminated by noise, which makes it very difficult to perform target recognition in SAR images. Inspired by great success of very deep convolutional neural networks (CNNs), this paper proposes a robust feature extraction method for SAR image target classification by adaptively fusing effective features from different CNN layers. First, YOLOv4 network is fine-tuned to detect the targets from the respective MF SAR target images. Second, a very deep CNN is trained from scratch on the moving and stationary target acquisition and recognition (MSTAR) database by using small filters throughout the whole net to reduce the speckle noise. Besides, using small-size convolution filters decreases the number of parameters in each layer and, therefore, reduces computation cost as the CNN goes deeper. The resulting CNN model is capable of extracting very deep features from the target images without performing any noise filtering or pre-processing techniques. Third, our approach proposes to use the multi-canonical correlation analysis (MCCA) to adaptively learn CNN features from different layers such that the resulting representations are highly linearly correlated and therefore can achieve better classification accuracy even if a simple linear support vector machine is used. Experimental results on the MSTAR dataset demonstrate that the proposed method outperforms the state-of-the-art methods.

In general, synthetic aperture radar (SAR) imaging and image processing are two sequential steps in SAR image processing. Due to the large size of SAR images, most image processing algorithms require image segmentation before processing. However, the existence of speckle noise in SAR images, as well as poor contrast and the uneven distribution of gray values in the same target, make SAR images difficult to segment. In order to facilitate the subsequent processing of SAR images, this paper proposes a new method that combines the back-projection algorithm (BPA) and a first-order gradient operator to enhance the edges of SAR images to overcome image segmentation problems. For complex-valued signals, the gradient operator was applied directly to the imaging process. The experimental results of simulated images and real images validate our proposed method. For the simulated scene, the supervised image segmentation evaluation indexes of our method have more than 1.18%, 11.2% and 11.72% improvement on probabilistic Rand index (PRI), variability index (VI), and global consistency error (GCE). The proposed imaging method will make SAR image segmentation and related applications easier.

This paper proposes a novel method on Synthetic Aperture Radar (SAR) Image denoising, which is based on context modeling combined with multiscale orthogonal bandlet coefficients. For the main influence of SAR image is multiplicative speckle noise, the logarithm uniform transform is used to convert it to additive noise. In this paper, the multiscale orthogonal bandlet transform is applied to noisy SAR Image. The contextual model is used to compute contextual values of bandlet coefficients on each scale. Therefore, bandlet coefficients are reconstructed into some parts by contextual values. The spatially adaptive soft-threshold (T) is to denoise SAR image in each part. Experimental results demonstrate that this method outperforms the methods of Ehance-Lee (E-lee), contextual model based on wavelet domain (Wavelet-context), and Hidden Markov Tree model bases on wavelet domain (Wavelet-hmt) on there indexes: Floating index (FI), Standard deviation and Edge saving index of vertical direction, by using a real SAR image. However, the method can not only reduce the speckle noise but also preserve edges and high frequency information, and gets considerable visual effect on SAR Image denoising.

The importance of synthetic aperture radar (SAR) images has increased exponentially in recent years due to its ability to capture images in all-weather, day and night conditions. SAR can generate target images in scale and cross-range with enough resolution for identification. The main demerit with SAR images is the multiplicative or speckle noise appearance in captured images, which can affect classification performance. To achieve a good classification result, it is crucial to reduce the speckle noise from the SAR imagery. The size of the available SAR imagery dataset is limited. There is always a chance that a classification model may overfit due to the small size dataset and less feature compared to optical imagery. In this work, the authors anticipated a new convolutional neural network (CNN) for SAR object classification based of despeckling. The experiment is performed on the MSTAR dataset. Experimental results of the current study showed that the proposed CNN model has resulted in an excellent training accuracy of 99.67 % and validation accuracy of 98.98%.

In single band and single polarized synthetic aperture radar (SAR) images, the information is limited to intensity and texture only and it is very difficult to interpret such SAR images without any a priori information. For unsupervised classification of SAR images, M-band wavelet decomposition is performed on the SAR image and sub-band selection on the basis of energy levels is applied to improve the classification results since sparse representation of sub-bands degrades the performance of classification. Then, textural features are obtained from selected sub-bands and integrated with intensity features. An adaptive neuro-fuzzy algorithm is used to improve computational efficiency by extracting significant features. K-means classification is performed on the extracted features and land features are labeled. This classification algorithm involves user defined parameters. To remove the user dependency and to obtain maximum achievable classification accuracy, an algorithm is developed in this paper for classification accuracy in terms of the parameters involved in the segmentation process. This is very helpful to develop the automated land-cover monitoring system with SAR, where optimized parameters are to be identified only once and these parameters can be applied to SAR imagery of the same scene obtained year after year. A single band, single polarized SAR image is classified into water, urban and vegetation areas using this method and overall classification accuracy is obtained in the range of 85.92%–93.70% by comparing with ground truth data.

In this paper, we propose a synthetic aperture radar (SAR) image change detection algorithm based on SAR image nonstationary analysis and conditional random field (CRF) model. This algorithm segments the SAR images into two nonstationary regions with the Triplet Markov Field (TMF) model which incorporates a U field to describe the images' nonstationary nature. The combined posterior probability of the CRF model consists of two parts: the unary potential and the pairwise potential. The unary potential is modeled by support vector machine (SVM) using the texture features extracted from the two-temporal SAR images, and the pairwise potential is constructed by the multilevel logistical model to capture the spatial interactions of the log-ratio image. Both the unary and pairwise potentials are improved with the introduction of the U field. The parameters of the pairwise potential are estimated by the iterative conditional estimation (ICE) algorithm. The iterative conditional model (ICM) algorithm is utilized for updating the combined posterior probability, and the final change detection results are obtained when it converges. Finally, experimental results on four two-temporal SAR images validate the effectiveness of the proposed image change detection algorithm.

The existing Synthetic Aperture Radar (SAR) image target detection methods based on convolutional neural networks (CNNs) have achieved remarkable performance, but these methods require a large number of target-level labeled training samples to train the network. Moreover, some clutter is very similar to targets in SAR images with complex scenes, making the target detection task very difficult. Therefore, a SAR target detection network based on a semi-supervised learning and attention mechanism is proposed in this paper. Since the image-level label simply marks whether the image contains the target of interest or not, which is easier to be labeled than the target-level label, the proposed method uses a small number of target-level labeled training samples and a large number of image-level labeled training samples to train the network with a semi-supervised learning algorithm. The proposed network consists of a detection branch and a scene recognition branch with a feature extraction module and an attention module shared between these two branches. The feature extraction module can extract the deep features of the input SAR images, and the attention module can guide the network to focus on the target of interest while suppressing the clutter. During the semi-supervised learning process, the target-level labeled training samples will pass through the detection branch, while the image-level labeled training samples will pass through the scene recognition branch. During the test process, considering the help of global scene information in SAR images for detection, a novel coarse-to-fine detection procedure is proposed. After the coarse scene recognition determining whether the input SAR image contains the target of interest or not, the fine target detection is performed on the image that may contain the target. The experimental results based on the measured SAR dataset demonstrate that the proposed method can achieve better performance than the existing methods.

ABSTRACTThe speckle noise of synthetic aperture radar (SAR) images limits its application in change detection. Compared with improved ratio (IR) and log-ratio (LR) operators, the neighborhood-based ratio (NR) technique can restrain the influence of speckle noise and is more suitable for change detection in SAR images. However, we find three drawbacks of NR by analyzing this method carefully. To overcome these defects, we propose an improved neighborhood-based ratio (INR) approach for change detection in SAR images. INR restructures the NR operator to exploit the neighborhood information more reasonably and is expected to reduce the impact of speckle noise better. IR, LR, mean ratio operator, NR, and INR are tested on two data sets to compare their performances in change detection of SAR images. Experimental results show that the proposed method can obtain better difference image than other state-of-art methods and improve the accuracy of change detection in SAR images effectively.

Synthetic aperture radar (SAR) images provide all-weather and all-time capabilities for Earth observation, which becomes highly beneficial in the field of intelligent remote sensing (RS) image interpretation. Due to these advantages, SAR images have been widely exploited in automatic building segmentation tasks under poor weather conditions, especially when disasters happen. However, compared to optical images, the semantics inherent to SAR images are less rich and interpretable due to factors such as speckle noise and imaging geometry. In this scenario, most state-of-the-art methods are focused on designing advanced network architectures or loss functions for building footprint extraction. However, few works have been oriented toward improving segmentation performance through knowledge transfer from optical images. In this article, we propose a novel method based on the <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DisOptNet</monospace> network, which can distill the useful semantic knowledge from optical images into a network only trained with SAR data. Specifically, we first analyze the multilevel feature discrepancies between multiple stages of the networks pretrained on the two image modalities. We observe that feature discrepancies start to increase as the encoding stage gradually changes from low level to high level. Based on such observation, we reuse the early stage features and construct parallel convolutional neural network (CNN) branches that are responsible for capturing high-level domain-specific knowledge for each image modality. The optical branch is aimed at mimicking feature generation at the optical pretrained network given the input SAR images. Then, an aggregation module is introduced to calibrate and fuse the features from different modalities while generating the building segments. Extensive experiments were conducted on a large-scale multisensor all-weather building segmentation dataset with state-of-the-art methods used for comparison. Our experimental results validate the effectiveness of <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DisOptNet</monospace> , which demonstrates great potential in the task of weather-independent building footprint generation under real scenarios. The codes of this article will be made publicly available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/jiankang1991/TGRS_DisOptNet</uri> .