Synthetic Aperture Radar Image Processing Research Articles

Theoretical Analysis of an Adaptive Periodic Multi Installment Scheduling With Result Retrieval for SAR Image Processing

Processing a large-scale Synthetic Aperture Radar (SAR) image dataset on a distributed computing infrastructure poses a challenging problem. Large-scale load distribution strategies like multi-installment scheduling (MIS) assume that the size of the result is negligible compared to the input workloads and hence ignore it in their design. Similarly, numerical methods like particle swarm optimization and their variants are not practical for real-time applications, given their run-time complexities. As both the results retrieval and completion time are crucial for SAR image data processing, in this article, we attempt to provide a thorough theoretical analysis of an adaptive MIS that includes the result retrieval phase. We use the periodic nature of the internal installments to keep the strategy simple and fine-tune the last installment to avoid any idle times in the processors. We derive a closed-form solution for the load fractions and hence, the overall processing time, schedule feasibility criteria, and certain other properties that lead to adaptive scheduling. Finally, we validate our theoretical findings through rigorous simulation studies using a loosely connected virtual machines (VMs) topology for the SAR dataset.

YOLO-SD: Small Ship Detection in SAR Images by Multi-Scale Convolution and Feature Transformer Module

As an outstanding method for ocean monitoring, synthetic aperture radar (SAR) has received much attention from scholars in recent years. With the rapid advances in the field of SAR technology and image processing, significant progress has also been made in ship detection in SAR images. When dealing with large-scale ships on a wide sea surface, most existing algorithms can achieve great detection results. However, small ships in SAR images contain little feature information. It is difficult to differentiate them from the background clutter, and there is the problem of a low detection rate and high false alarms. To improve the detection accuracy for small ships, we propose an efficient ship detection model based on YOLOX, named YOLO-Ship Detection (YOLO-SD). First, Multi-Scale Convolution (MSC) is proposed to fuse feature information at different scales so as to resolve the problem of unbalanced semantic information in the lower layer and improve the ability of feature extraction. Further, the Feature Transformer Module (FTM) is designed to capture global features and link them to the context for the purpose of optimizing high-layer semantic information and ultimately achieving excellent detection performance. A large number of experiments on the HRSID and LS-SSDD-v1.0 datasets show that YOLO-SD achieves a better detection performance than the baseline YOLOX. Compared with other excellent object detection models, YOLO-SD still has an edge in terms of overall performance.

SAR Antenna Pattern Measurement by Internal Calibration Method for GF-3 Satellite

GF-3 is the first C-band multipolarization synthetic aperture radar (SAR) satellite in China. The SAR system is equipped with an active phased array antenna to electronically generate various antenna beams. The accuracy of SAR antenna patterns is of main importance for precise SAR image processing. The traditional method for antenna pattern measurement is the near-field method but it is unable to satisfy the demand for fast measuring in-orbit. In order to measure the SAR antenna pattern quickly with high accuracy, we propose the internal calibration method. The internal calibration uses calibration signals which are routed along the nominal signal path in SAR systems, thus monitoring the gain and phase variations for each T/R channel is possible. This paper will focus on this method, including its principle, operation steps, and results, compared to the near-field method. The internal calibration method provides a valuable solution for SAR antenna pattern measurement.

Sparse SAR Imaging Method for Ground Moving Target via GMTSI-Net

Ground moving targets (GMT), due to the existence of velocity in range and azimuth direction, will lead to the deviation from their true position and defocus in the azimuth direction during the synthetic aperture radar (SAR) imaging process. To address this problem and compress the amount of echo data, a sparse SAR imaging method for ground moving targets is proposed. Specifically, we first constructed a two-dimensional sparse observation model of the GMT based on matched filter operators. Then, the observation model was solved by a deep network, GMT sparse imaging network (GMTSI-Net), which was mainly obtained by unfolding an iterative soft threshold algorithm (ISTA)-based iterative solution. Furthermore, we designed an adaptive unfolding module in the imaging network to improve the adaptability of the network to the input of echo data with different sampling ratios. The proposed imaging network can achieve faster and more accurate SAR images of ground moving targets under a low sampling ratio and signal-to-noise ratio (SNR). Simulated and measured data experiments were conducted to demonstrate the performance of imaging quality of the proposed method.

Self-calibrated dilated convolutional neural networks for SAR image despeckling

ABSTRACT Synthetic aperture radar (SAR) imaging process is essentially disturbed by speckle noise due to its unique mechanism. Speckle noise causes severe degradation of SAR image quality, which significantly limits its practical application. Recently, convolutional neural networks (CNNs) have indicated good potential for various image processing tasks. In this article, we propose a self-calibrated dilated convolutional neural network for SAR image despeckling called SAR-SCDCNN. The main body of SAR-SCDCNN is formed by seven self-calibrated blocks (SeCaBlock). Firstly, in each SeCaBlock, the input features are split into two branches: one represents the contextual features in the original space, and another represents those in the long-range space. Then, the down-up sampling operation and the convolutions with hybrid dilated rates are employed to enlarge the receptive field. The weights for calculating features themselves are adaptively extracted in the second branch. And then, two branch features are concatenated and fused through a convolution. Finally, a skip connection is used between the input and output of each SeCaBlock to give full play to the expressive power of the deep neural network and enhance training stability. Experiments on synthetic speckled and real SAR images are conducted to perform objective quantitative and subjective visual evaluations of image quality. Results show that our proposed method can effectively suppress speckle noise and adequately preserve detailed features.

Canopy Height Mapping by Sentinel 1 and 2 Satellite Images, Airborne LiDAR Data, and Machine Learning

Continuous mapping of vegetation height is critical for many forestry applications, such as planning vegetation management in power transmission line right-of-way. Satellite images from different sensors, including SAR (Synthetic Aperture Radar) from Sentinel 1 (S1) and multispectral from Sentinel 2 (S2), can be used for producing high-resolution vegetation height maps at a broad scale. The main objective of this study is to assess the potential of S1 and S2 satellite data, both in a single and a multisensor approach, for modeling canopy height in a transmission line right-of-way located in the Atlantic Forest of Paraná, Brazil. For integrating S1 and S2 data, we used three machine learning algorithms (LR: Linear Regression, CART: Classification and Regression Trees, and RF: Random Forest) and airborne LiDAR (Light Detection and Ranging) measurements as the reference height. The best models were obtained using the RF algorithm and 20 m resolution features from only S2 data (cross-validated RMSE of 4.92 m and R2 of 0.58) or multisensor data (cross-validated RMSE of 4.86 m and R2 of 0.60). Although the multisensor model presented the best performance, it was not statistically different from the single-S2 model. Thus, the use of only S2 to estimate canopy height has practical advantages, as it reduces the need to process SAR images and the uncertainties due to S1 noise or differences between the acquisition dates of S2 and S1.

A COMPARISON BETWEEN CYCLE-GAN BASED FEATURE TRANSLATION AND OPTICAL-SAR VEGETATION INDICES

Abstract. Optical and microwave remote sensing technologies have become key tools for local and global change detection applications. Generally, optical data has been the focus of remote sensing for change detection because of the varied spatial and temporal resolutions that allow for reliable information. However, the dependence of optical data on weather conditions prevents continuous and up-to-date information. On the other hand, Synthetic Aperture Radar (SAR) data can record all-weather and all-time polarization information which is critical for change detection in poor weather conditions; nonetheless, SAR is not precise as optical data for forestry change detection applications as it cannot provide the spectral features of interest. The combined processing of optical and SAR images allow for the retrieval of information of interest with a precision that none of them could achieve alone. In this context, Cycle-Consistent Generative Adversarial Networks (CycleGAN) based deep feature translation method was proposed in this study for change detection. The CycleGAN transfers images from one domain (optical) to another domain (SAR) into the same feature space using a cyclic structure. Thus, it can provide continuous and up-to-date information for change detection while keeping its spectral features. The accuracy of the fake images generated from CycleGAN was evaluated by correlating them with spectral indices (e.g., Normalized Difference Vegetation Index (NDVI), Modified Radar Vegetation Index (mRVI), and Modified Radar Forest Degradation Index (mRFDI)) directly obtained from optical and SAR data. As a result, the best correlation coefficients (R) were found between real NDVI (optical data) and fake NDVI (CycleGAN) with 0.98 and 0.97 for two different dated datasets.

Restoration of speckle noise corrupted SAR images using regularization by denoising

Speckle noise removal is a well-established problem in synthetic aperture radar (SAR) image processing. Among different methods focused on the reconstruction of SAR images, variational models have achieved state-of-the-art performance. In this paper, a Rayleigh based speckle reduction algorithm is developed using the variational framework. The forward model is combined with recently proposed regularization by denoising (RED) prior. However, RED has been proposed in literature for the additive noise model. Multiplicative noise in SAR images prevents the direct application of RED to variational models. Hence, logarithm transformation is applied to change the multiplicative noise model to additive model, and the forward model from Rayleigh to Fisher–Tippett distribution. The resulting optimization problem is solved using the alternating direction method of multipliers. Further, the proof of the convergence analysis is carried out for the above framework. Simulations convey that the proposed method has better despeckling performance compared to that of state-of-the-art methods.

Stabilized Detection Accuracy Maximization Using Adaptive SAR Image Processing in LEO Networks

The use of low Earth orbit (LEO) satellites for world-wide surveillance services is currently actively discussed and developed because the constellation of satellites is one major approach which can provide global seamless network services. Because synthetic aperture radar (SAR), which is used for satellite image acquisition and its related signal processing, is dealing with large volumes of image data, corresponding on-demand adaptive methods for SAR image processing are essentially required for stabilized surveillance services under the consideration of data burst situations. Thus, an adaptive vision algorithm for ship detection which is one of major tasks in SAR image processing researches is proposed based on Lyapunov optimization framework, which maximizes the detection performance while satisfying stability conditions. The high-performance filters are utilized for precisely recognizing the targets whereas they introduce relatively larger delays (i.e., tradeoff exists between performances and delays). Therefore, the proposed Lyapunov optimization-based adaptive filter selection algorithm is designed based on the characteristics. Our data-intensive performance evaluation results prove that the proposed algorithm achieves desired performance improvements.

Oriented Ship Detector for Remote Sensing Imagery Based on Pairwise Branch Detection Head and SAR Feature Enhancement

Recently, object detection in natural images has made a breakthrough, but it is still challenging in oriented ship detection for remote sensing imagery. Considering some limitations in this task, such as uncertain ship orientation, unspecific features for locating and classification in the complex optical environment, and multiplicative speckle interference of synthetic aperture radar (SAR), we propose an oriented ship detector based on the pairwise branch detection head and adaptive SAR feature enhancement. The details are as follows: (1) Firstly, the ships with arbitrary directions are described with a rotated ground truth, and an oriented region proposal network (ORPN) is designed to study the transformation from the horizontal region of interest to the rotated region of interest. The ORPN effectively improved the quality of the candidate area while only introducing a few parameters. (2) In view of the existing algorithms that tend to perform classification and regression prediction on the same output feature, this paper proposes a pairwise detection head (PBH) to design parallel branches to decouple classification and locating tasks, so that each branch can learn more task-specific features. (3) Inspired by the ratio-of-average detector in traditional SAR image processing, the SAR edge enhancement (SEE) module is proposed, which adaptively enhances edge pixels, and the threshold of the edge is learned by the channel-shared adaptive thresholds block. Experiments were carried out on both optical and SAR datasets. In the optical dataset, PBH combined with ORPN improved recall by 5.03%, and in the SAR dataset, the overall method achieved a maximum F1 score improvement of 6.07%; these results imply the validity of our method.

Design and Implementation of Synthetic Aperture Radar (SAR) Field-Programmable Gate Array (FPGA)-Based Processor

Synthetic aperture radar (SAR) is a unique imaging radar system that is capable of obtaining high-resolution images by using signal-processing techniques while operating in all weather and in the absence of a light source. The potential of SAR in a wide range of applications has led to new challenges in digital SAR processor design. On-board storage of SAR raw data is often not practical for real-time applications. The design of digital SAR processors is always restricted by the available space of the carrier system, data transfer rate, payload capacity and on-board power supplies. As reported in the literature, although customized hardware solutions could offer the desired performance, they are not feasible for low-volume production. This research aims to design and develop an efficient digital SAR processor by using field-programmable gate array (FPGA) with the consideration of hardware resources, processing speed and precision. In this paper, a hardware implementation of an FPGA-based SAR processor is presented. The implementation and architecture of the proposed SAR processor are highlighted in this paper. A MATLAB-based SAR processing range-Doppler algorithm (RDA) was developed as the benchmark for the development of an SAR processor. The target device, Altera Stratix IV GX FPGA EP4SGX230KF40C2, was selected for the design and implementation of an FPGA-based SAR processor. Comprehensive evaluations of the performance of the proposed SAR processor in terms of precision, timing performance and hardware resource utilizations are also presented. The proposed FPGA-based digital SAR processor achieves optimum performance in processing SAR signals for image formation. Evaluation shows that the designed SAR processor is capable of processing SAR images with ±1% difference error as compared to SAR images processed by MATLAB. The results also show a reduction in hardware usage via the implementation of an FPGA-based FFT/IFFT coprocessor. These promising results prove that the performance of the proposed processor is satisfactory and the achieved processing time, as well as the power consumption of the processor, outperformed existing implementations.

Bistatic SAR Point Spread Function Analysis for Close Proximity Geometries

The goal of this project is to describe the spatially variant imaging performance of bistatic SAR imaging using a localized spatially invariant approximation so as to use spatial Fourier analysis. Bistatic synthetic aperture radar (SAR) has been proposed for use in search and rescue operations, for UAV guidance, and as an all-weather navigational aid for landing aircraft. These applications involve aggressive maneuvering platforms operating at short standoff ranges. Matched filter image formation (MFIF) and the equivalent back-projection image formation algorithm fully capture the intrinsic imaging performance of these geometries. The imaging performance for these imaging configurations is described by the spatially variant point spread function (PSF). The SAR imaging process is described as a non-convolutional integral operation between the scene’s reflectivity function and the PSF. It is desirable to describe the process as a multiplication operation in the spatial Fourier domain but this requires the imaging process to be spatially invariant. The model derived here deals with the spatial variations by segmenting a large scene into smaller regions. Within each region, the spatially variant PSF is locally approximated as a spatially invariant impulse response (IPR). The SAR imaging process is described as a localized convolution operation. This approximation is extremely accurate at the center of a segmented region but degrades for positions displaced from the center. Error analysis is used to determine the maximum size of a segment to limit the error to a given tolerance.

Unsupervised Deep Sparse Features Extraction for SAR Image Segmentation

Deep learning (DL) methods usually need to collect a large amount of labeled data to extract deep features. However, due to the difficulty of obtaining numerous labeled data from synthetic aperture radar (SAR) images, unsupervised feature learning has been focused on SAR image processing. In this paper, we propose a three-dimensional sparse model (3-DSM) to extract deep sparse features from SAR images in an unsupervised way. Concretely, 3-DSM learns the convolution kernels by minimizing the error between the features and the constructed sparse maps, without labeled samples. Thus, the discriminative features can be extracted in an unsupervised way by the learned convolution kernels and are able to capture the main structure information of SAR images. Furthermore, to the best of our knowledge, 3-DSM firstly specifies the sparsity of convolution kernels, with each convolution kernel exhibiting its independence from the others and the redundancy of convolution kernels being diminishing. It means that each convolution kernel extracts its unique structural features of SAR images. Consequently, in the feature extraction, three-dimensional sparsities have been specified, including width, height, and depth, with the acquisition of discriminative less-redundant features. The effectiveness of 3-DSM is demonstrated by the feature extraction and segmentation of the simulated and real SAR images.

Scattering Model Guided Adversarial Examples for SAR Target Recognition: Attack and Defense

Deep Neural Networks (DNNs) based Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) systems have shown to be highly vulnerable to adversarial perturbations that are deliberately designed yet almost imperceptible but can bias DNN inference when added to targeted objects. This leads to serious safety concerns when applying DNNs to high-stakes SAR ATR applications. Therefore, enhancing the adversarial robustness of DNNs is essential for applying DNNs to modern real-world SAR ATR systems. Toward building more robust DNN-based SAR ATR models, this article explores the domain knowledge of SAR imaging process and proposes a novel Scattering Model Guided Adversarial Attack (SMGAA) algorithm which can generate adversarial perturbations in the form of electromagnetic scattering response (called adversarial scatterers). The proposed SMGAA consists of two parts: 1) a parametric scattering model and corresponding imaging method and 2) a customized gradient-based optimization algorithm. First, we introduce the effective Attributed Scattering Center Model (ASCM) and a general imaging method to describe the scattering behavior of typical geometric structures in the SAR imaging process. By further devising several strategies to take the domain knowledge of SAR target images into account and relax the greedy search procedure, the proposed method does not need to be prudentially finetuned, and can efficiently find the effective ASCM parameters to fool the SAR classifiers and facilitate the robust model training. Comprehensive evaluations on the MSTAR dataset show that the adversarial scatterers generated by SMGAA are more robust to perturbations and transformations in the SAR processing chain than the currently studied attacks, and are effective to construct a defensive model against the malicious scatterers.

A Benchmark Sentinel-1 SAR Dataset for Airport Detection

Airports are important targets in both military and civilian tasks. Synthetic aperture radar (SAR) based airport detection has received increasing attention in recent years. However, due to the high cost of the SAR imaging and annotation process, there is no publicly available SAR dataset for airport detection, which leads to the fact that deep learning methods have not been applied to airport detection and no unified evaluation benchmark for traditional methods. To provide a benchmark for airport detection studies in SAR images, a large SAR airport dataset (SAD) is presented in this paper. It contains 624 SAR images from Sentinel 1B, covering 104 instances of airports with different scales, orientations and shapes, which can realistically reflect the real world. Experiments with four deep learning-based methods and three traditional methods on this dataset demonstrate its effectiveness and challenge. It serves for the development of state-of-the-art airport detection algorithms or other related tasks. In addition, we found that airport runways in SAR images always have some parallel line segments. Inspired by this, the fusion of deep features and line segments is achieved by designing a line segment detector branch, which further improves the accuracy of airport detection.

Location of Synthetic Aperture Radar Imagery via Range History

Geolocation is one of the key synthetic aperture radar (SAR) image processing procedures in SAR-based remote sensing applications. The conventional SAR imagery geolocation model is based on the range-Doppler (RD) equation, which is derived from the Fresnel approximation of SAR. Currently, there is no other rigorous geolocation model for modern high-resolution SAR; however, this approximation may not be fully applicable for various modern SAR imaging modes and algorithms. This study aims to develop a novel SAR image geolocation model based on the SAR range history (RH) equation, which is derived from the basic SAR backscattered signal property. Moreover, a new geometric calibration model is presented for RH-based location. Experiments based on real spaceborne SAR data are presented as examples, and the location accuracy of RH and RD location models are compared. The results show that the location accuracies of the RH and RD models are consistent under the existing SAR processing framework, but those for the RH model demonstrate better performance in the azimuth direction. In theory, the proposed RH model can be used for SAR image products generated from complex modes and non-Fresnel-approximation-based imaging algorithms, making it suitable for future SAR systems.

Change Detection in SAR Images via Ratio-Based Gaussian Kernel and Nonlocal Theory

Compared with the synthetic aperture radar (SAR) image processing theory based on local neighborhood, the nonlocal theory is not limited to a local neighborhood of an image and has great potential in change detection of SAR images. In this study, an approach using ratio-based nonlocal information (RNLI) is proposed for change detection in multitemporal SAR images. First, the RNLI is extracted from a spatial–temporal nonlocal neighborhood where the similarity of two pixels in the nonlocal neighborhood is well characterized by the proposed ratio-based Gaussian kernel function. The parameters of RNLI: noise level and matching window size are adaptively determined to avoid the uncertainty of the change detection result caused by user experience. Second, the difference image is generated by using the RNLI and the ratio operator. Finally, the change map is obtained by segmenting the difference image with a threshold. Experiments conducted on two real datasets and two simulated datasets showed that the proposed method performed better than the other advanced change detection methods, which can better retain the edge information of the changed area while reducing the overall error of the change detection results.

Synthetic Aperture Radar Image Formation and Processing on an MPSoC

Satellite remote sensing acquisitions are usually processed after downlink to a ground station. The satellite travel time to the ground station adds to the total latency, increasing the time until a user can obtain the processing results. Performing the processing and information extraction onboard of the satellite can significantly reduce this time. In this study, synthetic aperture radar (SAR) image formation as well as ship detection and extreme weather detection were implemented in a multiprocessor system on a chip (MPSoC). Processing steps with high computational complexity were ported to run on the programmable logic (PL), achieving significant speed-up by implementing a high degree of parallelization and pipelining as well as efficient memory accesses. Steps with lower complexity run on the processing system (PS), allowing for higher flexibility and reducing the need for resources in the PL. The achieved processing times for an area covering 375 km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> were approximately 4 s for image formation, 16 s for ship detection, and 31 s for extreme weather detection. These developments combined with new downlink concepts for low-rate information data streams show that the provision of satellite remote sensing results to end users in less than 5 min after acquisition is possible using an adequately equipped satellite.

Fully Size-Adaptive Sliding Window Method for SAR Image Edge Detection

The local statistics are useful for the synthetic aperture radar (SAR) image processing. However, unreliable statistical results are produced due to insufficient sample size. To overcome the statistical limitation, the appropriate size of sliding window is vital in the image processing. Therefore, a completely size-adaptive sliding window method is proposed in this letter. A method to determine the size boundaries of the sliding window based on the basic size information is firstly proposed. Then, a strategy for changing the sliding window size is used to generate the sliding window size index matrix of the image. Among the many image processing applications that use sliding windows, we choose the SAR image edge detection arbitrarily to verify the effectiveness of the proposed method. And the experimental results indicate that our method makes the original edge detection algorithms improved significantly. Further, the proposed size-adaptive sliding window method has potential for many other image processing tasks that require the use of local statistical information.

An Efficient Radio Frequency Interference Mitigation Algorithm in Real Synthetic Aperture Radar Data

As a wideband radar system, a synthetic aperture radar (SAR) may conflict with several electromagnetic systems, such as frequency modulation (FM), TV, and other communication systems. These signals, termed as radio frequency interference (RFI), may severely interfere SAR systems from generating a high-resolution image. Numerous previous researches focused on the RFI suppression problem, among which the semiparametric methods have been verified to have the state-of-the-art performance. However, most of the semiparametric methods are computationally expensive and can hardly be used on wide-swath SAR imaging processing. In this article, an efficient semiparametric algorithm is proposed to suppress RFIs via alternating projections. It has comparable performance as the other methods but significantly improves the computational efficiency a lot. It is able to remove both narrowband and wideband RFIs and can be used directly on the Level-1 SAR data. Finally, multiple real SAR data are provided to demonstrate the effectiveness and efficiency of the proposed algorithm.