Automatic Synthetic Aperture Radar Research Articles

A Survey on SAR and Optical Satellite Image Registration

After decades of research, automatic synthetic aperture radar (SAR)-optical registration remains an unsolved problem. SAR and optical satellites utilize different imaging mechanisms, resulting in imagery with dissimilar heterogeneous characteristics. Transforming and translating these characteristics into a shared domain has been the main challenge in SAR-optical matching for many years. Combining the two sensors will improve the quality of existing and future remote sensing applications across multiple industries. Several approaches have emerged as promising candidates in the search for combining SAR and optical imagery. In addition, recent research has indicated that machine learning-based approaches have great potential for filling the information gap posed by utilizing only one sensor type in Earth observation applications. However, several challenges remain, and combining them is a multi-step process where no one-size-fits-all approach is available. This article reviews traditional, state-of-the-art, and recent development trends in SAR-optical co-registration methods.

Fast Superpixel-Based Clustering Algorithm for SAR Image Segmentation

In this letter, we propose a fast superpixel-based clustering algorithm (FSC) for synthetic aperture radar (SAR) image segmentation. First, the SAR image is over-segmented into superpixels by our previously proposed edge-aware superpixel generation method with one iteration merging (ESOM). Second, based on the obtained superpixels, the number of clusters is automatically selected by the density peak (DP) algorithm and knee point method instead of manual specification. Finally, the modified <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> -means clustering with the generalized-likelihood ratio (GLR) dissimilarity is performed on the superpixels to generate the final segmentation result. Experimental results on two real SAR images show that the proposed method outperforms other state-of-the-art methods in terms of both segmentation accuracy and computational efficiency. Moreover, our method is free of clustering parameters and achieves automatic SAR image segmentation.

Integrating the Reconstructed Scattering Center Feature Maps With Deep CNN Feature Maps for Automatic SAR Target Recognition

Automatic target recognition has been one of the hottest research in synthetic aperture radar (SAR) data processing. Noticing that popular recognition methods cannot utilize multiple features of SAR complex data, a method fused scattering center feature and deep convolutional neural network (CNN) feature is proposed in this letter. This method contains three key parts, namely, scattering center extraction and reconstruction block, CNN feature extraction block, and final feature fusion and classification block. In this process, the scattering center feature and CNN feature are fused at the level of feature maps, which retain the space information of 2-D feature maps. What is more, the proposed half end-to-end strategy realizes the automatic update of weighting parameters in feature extraction network and subnetwork, which promotes a better recognition efficiency. Experimental results on measured SAR data show that the proposed method can achieve better accuracy than other single feature-based methods and feature fusion methods.

A Feature Decomposition-Based Method for Automatic Ship Detection Crossing Different Satellite SAR Images

In the face of Synthetic Aperture Radar (SAR) image object detection with different distributions of training and test data, traditional supervised learning methods cannot achieve good detection performance. Domain adaptation (DA) method has been shown to have the ability to solve this problem, but existing DA object detection algorithms all use adversarial DA theory for the detection task, which is ineffective in solving object regression localization in the detection task. In this article, to better solve the above problem, an automatic SAR image ship detection method based on feature decomposition crossing different satellites is proposed. The feature extraction layer of backbone network is divided into low level and high level, where domain-invariant feature extractors are designed for the local features extracted from the low level and the global features extracted from the high level, respectively. We argue that the local and global features extracted from source domain and target domain contain domain-specific features (DSF) for adversarial DA and domain-invariant features (DIF) that contribute to object regression localization. Then, we decompose the local features and global features into DSF and DIF via vector decomposition method. For DSF counterpart, we introduce adversarial DA attention for feature alignment. DIF from the local features are fused into the backbone network for high-level global feature extraction. Finally, by using region proposal network and adversarial domain classifier, we can get the accurate bounding box and object class of SAR image objects. Extensive experiments prove that the proposed method outperforms state-of-the-art methods in terms of detection performance.

An Automatic Ship Detection Method Adapting to Different Satellites SAR Images With Feature Alignment and Compensation Loss

Traditional deep learning Synthetic Aperture Radar (SAR) image object detection methods fail to provide effective detection results when faced with SAR image datasets with different joint probability distributions obtained from multiple imaging satellites. In this article, an automatic SAR image object detection method based on domain adaptation is proposed to adapt to unlabeled target domain datasets acquired by different satellites. On the basis of introducing an adversarial domain adaptation learning strategy, we propose Adversarial Learning Attention (ALA) and Compensation Loss Module (CLM) on the baseline network. In ALA, considering the great difference in the scattering intensity of SAR images, the entropy vector can be used to distinguish the high-entropy and low-entropy regions among them and assign the corresponding weights, based on which the adversarial domain adaptation learning attention is proposed to achieve instance-level feature alignment and pixel-level feature alignment in source domain and target domain, respectively. In CLM, the domain alignment of pixel-level feature and instance-level feature of SAR image objects is first implemented, and then to make the feature alignment of both domains more accurate, better aggregation of proposals of different classes of prototype objects in the same domain is required, and further compensation loss is proposed to further restrict the prototype alignment of SAR object features in both domains. We conduct experiments on two SAR datasets obtained from different satellites whose results show the superiority of the proposed method over the state-of-the-art (SOTA) methods and the effectiveness of the proposed module in improving the detection accuracy.

Change Detection from SAR Images Based on Convolutional Neural Networks Guided by Saliency Enhancement

Change detection is an important task in identifying land cover change in different periods. In synthetic aperture radar (SAR) images, the inherent speckle noise leads to false changed points, and this affects the performance of change detection. To improve the accuracy of change detection, a novel automatic SAR image change detection algorithm based on saliency detection and convolutional-wavelet neural networks is proposed. The log-ratio operator is adopted to generate the difference image, and the speckle reducing anisotropic diffusion is used to enhance the original multitemporal SAR images and the difference image. To reduce the influence of speckle noise, the salient area that probably belongs to the changed object is obtained from the difference image. The saliency analysis step can remove small noise regions by thresholding the saliency map, and interest regions can be preserved. Then an enhanced difference image is generated by combing the binarized saliency map and two input images. A hierarchical fuzzy c-means model is applied to the enhanced difference image to classify pixels into the changed, unchanged, and intermediate regions. The convolutional-wavelet neural networks are used to generate the final change map. Experimental results on five SAR data sets indicated the proposed approach provided good performance in change detection compared to state-of-the-art relative techniques, and the values of the metrics computed by the proposed method caused significant improvement.

Imbalanced Learning-Based Automatic SAR Images Change Detection by Morphologically Supervised PCA-Net

Change detection is a quite challenging task due to the imbalance between unchanged and changed class. In addition, the traditional difference map generated by log-ratio is subject to the speckle, which will reduce the accuracy. In this letter, an imbalanced learning-based change detection is proposed based on PCA network (PCA-Net), where a supervised PCA-Net is designed to obtain the robust features directly from given multitemporal synthetic aperture radar (SAR) images instead of a difference map. Furthermore, to tackle with the imbalance between changed and unchanged classes, we propose a morphologically supervised learning method, where the knowledge in the pixels near the boundary between two classes is exploited to guide network training. Finally, our proposed PCA-Net can be trained by the data sets with available reference maps and applied to a new data set, which is quite practical in change detection projects. Our proposed method is verified on five sets of multiple temporal SAR images. It is demonstrated from the experiment results that with the knowledge in training samples from the boundary, the learned features benefit change detection and make the proposed method outperform than supervised methods trained by randomly drawing samples.

Automatic Image Contrast Enhancement for Small Ship Detection and Inspection Using RADARSAT-2 Synthetic Aperture Radar Data

This study devises an automatic synthetic aperture radar (SAR) image enhancement method for ship detection and inspection for installation in a near-real-time automatic high-speed processing system. The proposed method was examined in small ship inspection and detection over the Ieodo Ocean area off Korea using RADARSAR-2 HV-polarization data. The proposed method involves four steps. First, the SAR input data is converted into a highly compressed gray scale image, which enables both computer screen display and high-speed processing due to its light volume. Second, the overall contrast is adjusted by power-law scaling to strengthen the target discrimination, which is attenuated because of the inefficiency of one-sided intensity distribution. This additionally provides excellent target visibility. Third, the intensity of the area in which targets and clutter coexist is rescaled from 0 to 255 using min-max linear stretching. This suppresses background clutter and makes targets more easily distinguishable from the clutter. Lastly, the remaining clutter is successfully eliminated using a median filter. As a result, an output image is obtained that is close to binary data and enables ship detection using only simple global thresholding. Our ship detection results were compared with that of ships identified using an automatic identification system and those visible in high-precision images by visual inspection. We verified that our method offers a high detection rate for small ships and does not involve complexity in distribution assumption, filtering or thresholding. The potential of our method is confirmed as an automatic SAR enhancement method for near-real-time ship detection and inspection.

An automatic optical and SAR image registration method with iterative level set segmentation and SIFT

Although optical image registration methods have been successfully developed over the past decades, the registration of optical and synthetic aperture radar (SAR) images is still a challenging problem in remote sensing. Feature-based methods are considered to be more effective for multi-source image registration. However, almost all of these methods rely on the feature extraction algorithms. In this article, a simultaneous segmentation and feature-based registration method based on an iterative level set and scale-invariant feature transform (ILS-SIFT) is proposed. The core idea consists of three aspects: (1) an iterative procedure that combines image segmentation and matching is proposed to avoid registration failure caused by poor feature extraction; (2) a uniform level set segmentation model for optical and SAR images is presented to segment conjugate features; and (3) an improved SIFT algorithm is employed to determine whether the registration was successful. Experimental results have shown the effectiveness and universality of the proposed method.

Typhoon eye extraction with an automatic SAR image segmentation method

Spaceborne microwave synthetic aperture radar (SAR), with its high spatial resolution (10–100 m), large area coverage, and day/night imaging capability, has been used as an important tool for typhoon monitoring. Since the microwave signal can penetrate through clouds, SAR images reveal typhoon morphology at the sea surface. Within the region of a typhoon eye, wind speed and the associated sea surface roughness are usually low. Therefore, the typhoon eye can be well distinguished as dark areas in SAR images. However, automatic typhoon eye extraction from SAR images is hampered by SAR image speckle noise and other false-alarm dark features contained in an image. In this study, we propose an image processing approach to extract typhoon eyes from SAR images. The three-step image processing includes: (1) applying an extended non-local means image denoizing algorithm to reduce image speckle noise; (2) applying a top-hat transform to denoized imagery to enhance the contrast; and (3) using a labelled watershed to segment the typhoon eye. Experimental results from analysing three Environmental Satellite SAR typhoon images show that our approach provides fast and efficient SAR image segmentation for typhoon eye extraction. Typhoon eyes are segmented correctly, and their edges are well detected. Our experimental results are comparable to manually extracted typhoon eye information. Fine-tuning of this approach will provide an automatic tool for typhoon eye information extraction from SAR images.

Change detection of boreal forest using bi-temporal ALOS PALSAR backscatter data

Long-wavelength Synthetic Aperture Radar (SAR) satellite systems have the potential to increase the efficiency of forest mapping and monitoring, which today are based primarily on optical satellite systems. Here, we evaluate the effectiveness of using L-band SAR satellite images to detect and delineate clear-cuts in Swedish boreal forest. A set of computationally efficient techniques are combined in a fully automated unsupervised bi-temporal change detection approach that detects changes in SAR backscatter intensities. For radiometric normalization and initial change classification, we propose an iterative procedure consisting of successive polynomial based histogram matching and thresholding. Recently proposed methods for automatic SAR amplitude ratio thresholding and final change classification are adopted. The latter is a Markov random field based change detection method that exploits both spectral and spatial information from one or multiple SAR polarization channels. The change detection approach was applied to SAR images from the Japanese Advanced Land Observing Satellite (ALOS) Phased Array type L-band Synthetic Aperture Radar (PALSAR) acquired in Fine Beam Dual (FBD) mode (HH- and HV-polarizations) with a pixel size of 20m (path data). Clear-cuts that took place between image acquisitions were clearly detected, and most errors were due to imperfect delineations of clear-cut edges. Pixel-wise clear-cut detection accuracies above 90% could be reached, with false alarm rates of approximately 10% or less. The results indicate that ALOS PALSAR path data are well suited for operational clear-cut detection in Swedish boreal forest.

On the SAR change detection review and optimal decision

Change detection is the process of identifying differences in the state of an object or scene after the occurence of an event. In this paper, we will present several similarity measures for automatic synthetic aperture radar (SAR) change detection, which can be classified into two families: the first regroups the measures based on pixel radiometry and the second collects the detectors based on local statistics. This article details a judicious method for SAR image change detection using Dezert–Smarandache Theory (DSmT). On the one hand, a Rayleigh distribution function is used to characterize globally the radar texture data, which allows mass assignment through the Kullback–Leibler distance. On the other hand, local pixel measurements are introduced through the Rayleigh Distribution Ratio indicator to refine the mass attribution and take into account the context information. Finally, DSmT is carried out by comparing the modelling results between temporal images. This technique has been applied on both simulated and real data and allows very satisfactory change detection results.

Statistical target behaviour in SAR images

The paper discusses the statistical behaviour of targets in synthetic aperture radar (SAR) images and demonstrates how this information can be exploited for automatic SAR target classification and recognition. It describes a target-smoothing algorithm for estimating the underlying radar cross-section, which can be used with an appropriate noise model to generate many target realisations for a given imaging configuration. This allows a statistical analysis of target features to be performed that can be used to enhance feature-based classification, as is demonstrated by a simple classification experiment using real targets.