Synthetic Aperture Radar (SAR) images are crucial for maritime vessel detection; however, challenges such as blurred ship edges, strong land scattering interference, and angular regression mismatches across varying target sizes hinder accurate rotational localization. In this paper, an oriented decoupling target detection method (R-MCLST) is proposed to address these issues. The method integrates three key modules: a multi-channel positioning module (MC-PM) that employs distributed average pooling and additional coordinate channels to enhance orientation awareness; a soft threshold-based multilayer perceptron (ST-MLP) that effectively mitigates background interference while robustly extracting complex features; and a Gaussian distribution-based prediction box (GD-BPB) that transforms rotated bounding box encoding into a two-dimensional Gaussian distribution using KL divergence for adaptive parameter adjustment. Experimental evaluations on the R-SSDD and MR-HRSID datasets demonstrate that R-MCLST achieves superior performance, with the R-SSDD dataset yielding AP50 of 87.48%, AP75 of 34.96%, AP of 41.15%, and AR of 46.52%, and the MR-HRSID dataset yielding AP50 of 61.59%, AP75 of 4.96%, AP of 19.13%, and AR of 22.24%. Comparative analyses confirm that the proposed method outperforms current state-of-the-art networks in accurately localizing rotating targets under challenging SAR imaging conditions.

Ship detection in Synthetic Aperture Radar (SAR) imagery has achieved significant progress but still faces several challenges. For example, single-modal SAR provides limited ship information, and identifying small ships in complex environments is difficult, affecting detection robustness. To address these issues, this letter proposes an efficient small ship detection network based on multi-modal fusion with unpaired SAR and AIS data, named MU-Net. To address the limitations of insufficient single-modal features, it is the first attempt to integrate SAR imagery with ship distribution density information from long-term AIS. Specifically, an Asymmetric Fusion Module (AFM) is designed to incorporate AIS spatial information into SAR imagery through weighted fusion, effectively overcoming the strict data pairing requirements of traditional multi-modal fusion methods and enhancing the spatial and texture feature representation of ships. Furthermore, to boost the performance of small ship detection in challenging scenarios, MU-Net designs a Feature Dynamic Optimization Module (FDOM) and a Channel Shuffle Module (CSM). FDOM adaptively adjusts convolutional filters, and CSM performs channel reconstruction, enhancing the multi-modal feature representation of small ships and effectively reducing interference from complex backgrounds. With the integration of multi-modal information, MU-Net achieves higher detection accuracy using a single detection head. Experimental results indicate that the proposed method outperforms state-of-the-art models, achieving 89.81% Precision, 90.47% Recall, 94.82% AP0.5, and 54.75 AP0.5:0.95, with only 5.25M parameters and 55.9 GFLOPs. The dataset and code will be available at: https://github.com/DLRS09/MU-Net.

Ship detection in remote sensing images has been attracting a lot of attention due to its great application value in both military and civilian fields. However, ships in high-resolution remote sensing images are characterized by the remarkable features of multiscale, arbitrary orientation, and dense arrangement, which is a great challenge for fast and accurate target detection. In order to solve problems, we propose a YOLOV5-based oriented ship detection method of remote sensing images with contextual global attention mechanism and lightweight task-specific context decoupling (CGTC-RYOLO) in this article. First, a cross-stage partial context transformer (CSP-COT) module is introduced to capture global contextual spatial relations using multihead self-attention (MHSA) to verify their implications in implicit dependencies. Second, we propose an angle classification prediction branch in the YOLOV5 head network for detecting targets in any direction and design probability and distribution loss function (PrfoIoU) to optimize the regression effect. Third, the lightweight task-specific context decoupling (LTSCODE) for target detection is employed to replace the original head in the YOLOV5 model, which is used to solve the accuracy problem caused by YOLOV5’s hybridization of classification and localization. Ablation experiments demonstrate the importance and effectiveness of each module. Compared with the benchmark model, the CGTC-RYOLO has the 5.9%, 3.7%, and 4.3% mAP improvements on the DOTA-ship dataset, the HRSC2016 dataset, and the UCAS-AOD dataset, respectively. Moreover, the model’s generalization is also validated. Compared with state-of-the-artmethods, the CGTC-RYOLO can achieve better accuracy and fewer parameters.

Synthetic aperture radar (SAR) and optical images possess complementary strengths, offering rich spatial and spectral information. The intelligent classification of features through image fusion of SAR and optical presents both opportunities and challenges. However, fusion and intelligent classification encounter hurdles. Different physical properties and imaging principles between SAR and optical images often lead to sensor property mismatches, causing information loss. Moreover, optical images are susceptible to weather conditions, while SAR images suffer from scattering noise interference. In addition, the nonuniform distribution of feature categories results in sample imbalance. To address these problems, this article proposed a new fusion network structure dual-encoder net (DEN). First, without increasing the model complexity, considering the differences in the performance of features under different sensors, this network was keyed to a composition of two encoders that were able to utilize their respective features to encode and reduce the impact of modal differences. Second, a detail attention module (DAM) was constructed to capture the detailed information that was obscured by the optical image and acquired by the SAR image. Finally, a new loss function, comprising weighted information loss, pixel loss, and noise loss, was introduced to mitigate sample imbalance, retain key information, and reduce noise effects. The experimental results showed that the proposed method outperforms the current popular image fusion methods, and the model complexity was improved by 15.3% while the overall accuracy (OA) was improved by 2.6%, and the entropy, peak signal-to-noise ratio (PSNR), and mean square error (MSE) were improved by 29%, 27%, and 7%, respectively.

Ship classification in synthetic aperture radar (SAR) images is a challenge in the field of ocean monitoring. On the one hand, there are few labeled samples in SAR remote sensing ship datasets, and a commonly used single classification criterion cannot effectively represent the distribution of categories. On the other hand, the small size of the SAR ship and the inconspicuous appearance characteristics lead to the fact that the SAR ship samples are with less discriminative information; therefore, the rich feature space of a ship cannot be effectively obtained, which increases the difficulty of target distinguishability. A multibranch embedding network with bi-classifier (MBEN-BC) model was proposed to address these problems and for few-shot SAR ship classification. First, the MBEN module was utilized to extract the multiscale feature map spatial information of the input image at multiple levels and establish cross-channel information interaction so as to obtain discriminative features at the local and global levels, which effectively enriched the feature space. Then, the BC module was constructed to represent the image features from the image level and descriptor level, respectively, and the two classification criteria were presented to promote a more compact distribution of similar samples in the feature space in order to effectively represent the distribution of categories with a small number of labeled samples. Experimental validation was carried out using the FUSAR-Ship, Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset, and OPENSAR-Ship dataset, and the MBEN-BC method achieved superior performance and good generalization ability compared to the current popular and state-of-the-art few-shot methods.

The change detection aims to identify temporal changes in land cover. In emergency disaster scenarios, acquiring postchange optical images is often difficult due to factors such as adverse weather and illumination conditions. In contrast, the SAR-based change detection is robust to these environmental factors but is prone to speckle noise and often lacks clear semantic interpretation. These challenges highlight the importance of multimodal approaches that integrate the complementary information from different data sources. To address the domain gap between optical and SAR data, we propose change detection prompt (CDPrompt), an automatic prompt-learning framework that leverages in-domain change information as prompts to suppress fake changes caused by the domain gap between the two modalities. CDPrompt incorporates a modality-specific domain tuning module (DTM) to inject the domain knowledge into the segment anything model (SAM), enabling efficient adaptation to multimodal data with minimal labels and training costs. A low-level enhancement module (LwEM) further refines spatial details using historical optical images, while a consistency loss enhances the learning of domain-invariant features between prechange optical and SAR images. To support evaluation in disaster scenarios with missing modalities, we extend the DFC25 dataset and introduce the first disaster-oriented multimodal change detection dataset, DFC25-Extended, comprising DFC25-OS-S and DFC25-O-SO. Extensive experiments on the Onera Satellite Change Detection (OSCD) and DFC25-Extended datasets demonstrate the superior performance and practical value of CDPrompt. The code and dataset will be publicly available at:https://github.com/zhanglimeng13/CDPrompt

Ocean wave is of great significance in marine environment prediction, maritime navigation, and global climate change. Synthetic aperture radar (SAR) is widely used in ocean wave spectra retrieval due to its 2-D high resolution, all-weather, and all-time advantages. Nevertheless, the nonlinear mapping between SAR and ocean waves, caused by velocity bunching, hinders the advancement of wave spectra inversion techniques, resulting in low-quality and incomplete wave spectra. To overcome the problem, a novel deep learning model SAR2WV for ocean wave spectra retrieval based on Pix2pix is proposed by constructing the nonlinear mapping relationship of SAR cross spectra and ocean wave spectra. A total of 106 844 Sentinel-1 wave mode dataset along with the corresponding European Centre for Medium-Range Weather Forecasts (ECMWF) ERA 5 wave data is processed and used for training the SAR2WV model. Experiments demonstrate that the proposed SAR2WV model can significantly improve the accuracy of the retrieved wave spectra and wave parameters, with the spectra similarity improved by 60.3%, root-mean-square error (RMSE) of significant wave height (SWH) decreased from 0.966 to 0.386 m, RMSE of mean wave period (MWP) decreased from 1.208 s to 0.811 s, and correlation coefficient of peak wave direction increased from 0.65 to 0.72, which achieves better performance than ocean swell wave spectra (OSW) algorithm and other methods.

Sea surface temperature (SST) is one of the most important parameters in the global ocean-atmosphere system. Predicting SST can help to analyze and identify extreme weather and protect marine environment in advance. Traditional numerical and machine learning methods tend to ignore spatial features. The single model in existing deep learning methods suffers from weakening spatial features and reducing ability of discriminating time-series information. At the same time, rare consideration about the influence of ocean physical phenomena has been given. These will lead to inaccurate prediction results. Based on the spatial-temporal characteristics and physical laws of the SST field, this paper proposes a hybrid multi-scale spatial-temporal model combining error correction map (ECM-HMSTM) to predict the SST. First, the ECM-HMSTM can comprehensively extract the spatial-temporal features of the SST field at different scales and thus the SST prediction map can be obtained. Second, by a new error correction approach based on the activity of tropical instability waves (TIWs), the ECM-HMSTM can effectively predict the variation characteristics of TIWs signals, which results in producing the error correction maps. Third, by fusing the two above maps, the SST field in the tropical eastern Pacific Ocean after five days is predicted. Experiment results show that the accuracy of the ECM-HMSTM was improved by 10.3% compared with the current state-of-the-art deep convolution model. Moreover, the SST predicted by the ECM-HMSTM performs well on characterizing the intensity of TIWs. Therefore, this paper provides a strategy for effective short-term prediction of SST fields, which is of guidance for prediction and analysis of ocean phenomena and climate.

This paper proposes a hierarchical double Dirichlet process mixture model (DDPMM) for multilook polarimetric synthetic aperture radar (PolSAR) data unsupervised classification. Specifically, within the framework of product model (PM), an observed PolSAR data point can be factorized as the multiplication representation of a positive-scalar texture variable and a complex-Wishart-distributed speckle component. Based on this assumption, the polarization DPMM and texture DPMM in the proposed model are hierarchically established to characterize the polarimetric matrix and texture variable, respectively, thus yielding the generation procedure of the observation data to be learned sufficiently. Meanwhile, instead of sharing the same texture vector in many existing PM-based methods, each data point in DDPMM is associated with its own texture vector, which can be characterized as the weighted summation of several densities via texture DPMM rather than following a single distribution, such that the texture information can be fully and flexibly captured. In particular, dual local spatial constraints based on the statistical representations of polarization and texture spaces are also explored on the two DPMMs, allowing the local correlation to be adequately and dynamically incorporated. Moreover, all closed-form updates are derived with the variational Bayesian inference algorithm and the cluster number of the proposed model can be determined automatically. Experimental results on four real PolSAR datasets demonstrate the superiority of the proposed DDPMM to some state-of-the-art methods.

Due to the confusion of target-like sparse structures and the interference of linear features in complex scenarios, many infrared small target detection methods struggle to effectively detect dim and small targets. In response to this challenge, we propose a new 3-D paradigm framework, which combines spatial-temporal weighting and regularization within a low-rank sparse tensor decomposition model. First, we design a novel spatial-temporal local prior structure tensor, named 3DST, which can significantly distinguish between targets and target-like sparse structures. Second, we introduce a three-directional log-based tensor nuclear norm (3DLogTNN) to provide a full characterization of the low-rankness of the background tensor. Third, we suggest a weighted three-directional total variation (3DTV) regularization to constrain smoothness features in background images. Finally, we develop an efficient alternating direction method of multipliers (ADMMs) to solve the proposed model. In particular, we devise a fast and accurate Sylvester tensor equation for accelerated subproblem solving. Extensive experimental results demonstrate that the proposed model has superior target detection and background suppression performance in complex scenarios compared with other detection methods.

This paper introduces an innovative approach in Synthetic Aperture Radar (SAR) polarimetry and proposes a novel descriptor called polarimetric autocorrelation matrix. Different from polarimetric covariance and coherency matrices, the polarimetric autocorrelation matrix can capture hidden Doppler information in the frequency domain and encode it in the phase using higher-order statistical methods. This matrix facilitates the joint extraction and analysis of polarization and Doppler information from fully Polarimetric SAR (PolSAR) data through matrix analysis. The paper explains that the polarimetric autocorrelation matrix can be decomposed into a Doppler-related matrix and a covariance-related matrix. The magnitudes and phases of these matrices provide insight into Doppler shifts between different polarizations and the variance of the backscattering coefficient. Our research demonstrates how the Doppler shift is influenced not only by the target’s radial velocity but also by radar polarization. Four quad-polarimetric RADARSAT-2 images are used for testing in this study, and a set of characterization features and parameters are derived from the polarimetric autocorrelation matrix. The novel descriptor, together with the new parameters, can detect man-made target and sea ice, mitigate ambiguity caused by moving targets, and indicate the motion status of ship targets and sea currents. Specifically for marine scenes, our data found Doppler differences in HH polarization and VV polarization induced by sea surface motion can be up to 100 Hz.

The detection of ships encompasses an abundance of applications within the domains of fishery management, marine rescue operations, and maritime monitoring. In recent years, a multitude of detectors based on deep learning have been utilized for the purpose of ship detection using synthetic aperture radar (SAR) images. However, disturbed by the strong scattering background on land and the influence of the SAR target scale, the existing detectors face great challenges in detecting inshore small ships. To solve this problem, this article proposes an oriented ship detection network for SAR images based on soft threshold and context information. First, a soft-threshold quantization module (STQM) based on the soft threshold function is designed to alleviate the interference of background noise on the feature map. Second, a local and global context fusion module (LGCFM) is designed to capture the contextual information of the target to enhance the detection of small targets. Third, the inclusion of the center loss in the loss function serves to impose additional constraints on the center coordinates and shape of the oriented bounding box. This is done to achieve a more balanced distribution of loss contributions across the various variables and to mitigate the target’s susceptibility to variations in the shape of the ground-truth bounding box. Finally, the proposed network is tested on the publicly available mini-rotated-high-resolution SAR images dataset (MR-HRSID) and Rotated SAR Ship Detection Dataset (R-SSDD) datasets. The results from experiments demonstrate that our approach attains state-of-the-art detection capabilities for inshore ships and small targets, while also effectively mitigating interference from terrestrial noises.

This letter proposes three new discriminators for dominant scattering mechanism (surface scattering or dihedral scattering) for the decomposition process of the compact polarimetric (CP) synthetic aperture radar (SAR) scattering model. Based on the covariance matrix of CP SAR under symmetry reflection assumption, aRe(SHHS*VV) reconstruction method based on the estimation of cross-polarization term is proposed, and three new discriminators,DISCREntropy,DISCREigen, andDISCRDoP, are obtained accordingly. Comparative experiments using real circular transmit linear receive (CTLR) CP SAR data for the three proposed discriminators as well as the commonly used discriminator g3demonstrate that the comprehensive decomposition results using the proposed three discriminators are better than the result using g3, andDISCRDoPis better thanDISCREigen,DISCREigenis better thanDISCREntropy.

Thanks to the powerful feature extraction and expression ability of convolutional neural networks (CNNs), exceptional success has been achieved in the field of ship automatic target recognition (ATR) of synthetic aperture radar (SAR). However, the CNNs cannot work effectively with sparse labelled samples and imbalanced categories.This study proposes a new Attention-Dense-CycleGAN (ADCG) method that is suitable for the ship transfer learning task from optical to SAR (OPT2SAR). The key improvement of the ADCG lies in the construction of a Dense Connection Module (DCM) and a lightweight Convolutional Block Attention Module (CBAM). The DCM is able to overcome the problems of generator feature redundancy, large network model parameters, and severe training time in the original CycleGAN network. The lightweight CBAM can solve the problem of not being able to locate the main features of ships with a minimal increase in network parameters. Compared with the performance of other popular generative adversarial networks, the superior performance of the ADCG in the OPT2SAR transfer learning is demonstrated with the Fréchet Inception Distance (FID) minimum of 76.04 and the Kernel Inception Distance (KID) minimum of 0.0403. Finally, the ability of pseudo-SAR domain images were tested to improve the recognition accuracy of popular ship classification networks, this achieved an average improvement of 6% in recognition accuracy. Therefore the results of this study verifies the rationality, validity, and application value of pseudo-SAR domain in solving the problems of sparse marker samples and class imbalance in ship ATR network model.

Fine-grained ship classification in optical remote sensing images is a major challenge in the ocean observation field, elaborated as follows: First, the cost of acquiring ship images is expensive. Obtaining numerous labeled samples is difficult, resulting in the poor generalization ability of training models. Second, the features of ship target cannot be accurately obtained owing to complex background interference. Third, inter-class similarity and intra-class diversity among different ships render ship classification difficult. In this study, we propose LATA-BP-BDC: a bi-prototype Brownian Distance Covariance (BDC) metric network with lightweight adaptive task attention (LATA) for few-shot fine-grained ship classification. First, the LATA module is used to generate 3-dimensional (3D) weights, which can effectively reduce complex background interference and improve the adaptive capturing ability of target features without including additional network operators. Second, we input target features into the BDC metric module and output the BDC matrices to represent image information. Because the similarity between two images can be calculated as the corresponding BDC matrices distance, the improvement of the relevance of similar targets can be realized. Finally, we use the bi-prototype module to generate highly accurate prototypes, further calibrating information differences between images, which enhances the correlation between the same category samples and separability between different categories samples. Consequently, this process effectively reduces the influence of large intra-class appearance variation and small inter-class appearance variation. We perform validation using two fine-grained datasets, FGSCR and CUB. Compared with state-of-the-art methods, the LATA-BP-BDC achieves a superior performance and has good generalization for fine-grained few-shot classification.

Model-based decomposition methods are widely used in full-polarization synthetic aperture radar (SAR), for the inversion and interpretation of ground features and constitute an important approach for understanding the behavior of backscattering. However, owing to the substantial differences between land and marine environments, different man-made and natural vegetation scattering structures render existing decomposition models unable to reasonably characterize scatterers on ships. Moreover, the combination of polarization decomposition models and neural networks for ship detection has rarely been investigated. Therefore, this study proposes a four-component decomposition model (Ship-4SD) suitable for describing the scattering characteristics of ships based on the surface scattering, double-bounce scattering, ±45° oriented dipole, and asymmetric scattering components. Furthermore, based on the differences in the scattering properties exhibited by different scattering components in ships and the powerful feature extraction capability of convolutional neural networks (CNNs), a scattering characteristic-aware fully polarized SAR ship detection network (SCANet) was designed to make full use of the scattering components in the decomposition model. Finally, the experimental results on a large amount of GF-3 fully polarized SAR data validated that the reasonability and superiority of Ship-4SD and SCANet. The Ship-4SD can better distinguish ship and clutter pixels compared to other four-component models and has a higher target-clutter ratio with respect to the multi-component models. SCANet proposed in this paper achieved an average precision of 94.43% and 96.56% on the GF-3 and SSDD datasets, respectively, which is better than that of other competitive CNN algorithms.

With the development of deep learning (DL), research on ship classification in synthetic aperture radar (SAR) images has made remarkable progress. However, such research has primarily focused on classifying large ships with distinct features, such as cargo ships, containers, and tankers. The classification of SAR fishing vessels is extremely challenging because of two main reasons: 1) the small size and minor interclass differences of fishing vessels make learning fine-grained features difficult, and 2) determining fishing vessel types is difficult, resulting in a lack of labeled data. Hence, after designing a process framework for vessel tagging, we construct a high-resolution fine-grained fishing vessel classification dataset (FishingVesselSAR), which contains 116 gillnetters, 72 seiners, and 181 trawlers. We then propose a novel DL model (FishNet) that aims to strengthen feature extraction and utilization. In FishNet, we introduce four innovative modules to ensure superior performance in SAR fishing vessel classification: a multipath feature extraction (MUL) module, a feature fusion (FF) module, a multilevel feature aggregation (MFA) module and a parallel channel and spatial attention (PCSA) module. Furthermore, we design an adaptive loss function to achieve better classification performance by mitigating the effects of class imbalance. In this paper, we report extensive ablation studies conducted to confirm the efficacy of the five improvements listed above. Sufficient comparisons with 33 advanced methods from the DL and SAR target classification communities demonstrate that FishNet achieves a SAR fishing vessel classification accuracy of 89.79%, which is 6.77% higher than that of the second-best method.

A challenging issue in the field of marine remote sensing is the application of polarimetric synthetic aperture radar (PolSAR) to small ship detection in complicated environments. Several outstanding polarimetric detectors (such as the optimal polarimetric detector, polarimetric whitening filter, and polarimetric notch filter, etc.), have been effectively implemented in practical applications. A linear combination model based on quadratic optimization is summarized to establish a general framework for polarimetric detectors, transitioning the PolSAR ship target detection from a model driven approach to a hybrid (model/data)-driven approach. However, the dimension of the covariance matrix may be high, and the computation cost will be large. The higher dimension of the covariance matrix requires a bigger the data demand. As a result, when the sample size is small, the model performance will degrade. In this paper, to decrease the computational complexity and improve the robustness, we propose a novel method called the simultaneous diagonalization transform (SDT). The proposed method enables an almost simplest representation of information from the covariance matrix providing a rapid detection algorithm. The simulation experiments demonstrate that polarimetric detectors based on SDT consistently outperform those based on other methods in terms of accuracy, efficiency, and sample size requirements across various complex backgrounds. Furthermore, the effectiveness, robust, and fastness of the polarimetric detector based on SDT is validated using real data collected by RadarSAT-2, GaoFen-3, and Sentinel-1A.

The polarimetric whitening filter (PWF) is widely used in constant false alarm rate (CFAR) ship detection in polarimetric synthetic aperture radar (PolSAR) imagery. The detection threshold plays a key role in the CFAR detection, which is generally determined by the statistical model of clutter. Many product models with different distributed textures are considered to fit the PWF output for an accurate threshold. Unfortunately, the product model with the general inverse Gaussian (G) texture, which can be namedG-Wishart model according to the polarimetric covariance matrix, has not been well studied for its effective calculation. In this letter, the probability density function (PDF), the probability of false alarm (PFA), and the threshold in the CFAR algorithm based on the PWF are all obtained corresponding to theGdistribution. The closed forms for the PDF and PFA are obtained with the Fox H function and its multivariate version, respectively. The threshold is derived by the bisection method when the false alarm rate (FAR) is constant. Finally, experimental results using both simulated and real data demonstrate that the different statistical models with the same log-cumulants can achieve almost the same CFAR loss.

The azimuth cutoff wavelength of SAR is an important parameter for retrieval of sea surface wind and wave. Earlier studies have fully demonstrated the substantial dependence of azimuth cutoff wavelength on polarization, but the present studies only focus on H-V linear polarization bases (HH, HV/VH, and VV) without considering the effects of other polarization bases (e.g., linear rotated, circular, and elliptical polarization). Benefiting from the quad-polarization advantage of GaoFen-3 SAR wave mode data and the support of polarization basis transformation theory, this study used 4,648 SAR data to study the correlation between cutoff wavelength and wind and wave parameters (e.g., significant wave height, and wind speed) under different polarization bases, and analyzed the variation of correlation coefficient caused by polarization basis change. Finally, the results were applied to evaluating the performance of wind and wave parameters retrieval. The results of the study show that the azimuth cutoff is strongly dependent on the polarization state of electromagnetic wave. The azimuth cutoff wavelength under the elliptical polarization bases has higher correlation with wind and wave than that under H-V linear, circular, and linear rotated polarization bases. Using the azimuth cutoff wavelength of the elliptical polarization bases can significantly improve the retrieval accuracy of wind and wave parameters. This study shall enhance the capabilities of polarized SAR systems to precisely derive more ocean surface properties. The result implies that polarization basis is an important factor that must be considered in future ocean SAR studies.

Ship detection via synthetic aperture radar (SAR) has been demonstrated to be very useful as polarimetric information helps discriminate between targets and sea clutter. Among the available polarimetric detectors, optimal polarimetric detection (OPD) theoretically provides the best detection performance under the assumption that the fully developed speckle hypothesis stands. This study proposes a polarimetric detection optimization filter (PDOF). The target clutter ratio (TCR) over the speckle variation was maximized using a matrix transform to derive the PDOF. The objective function based on a matrix transform instead of a vector transform is optimized to obtain synthetic effects by combining a polarimetric whitening filter (PWF) and a polarimetric matched filter (PMF). Subspace form of the PDOF (SPDOF) is also proposed, which gives performance comparable to the PDOF. Assuming a Wishart distribution, the exact and approximate expressions of the closed-form probability density function (PDF) of the PDOF are derived. The probability of false alarm (PFA) was derived in a closed-form expression, which allows obtaining the PDOF threshold analytically. Moreover, the gamma model is extended to a generalized gamma distribution ($\text{G}\Gamma \text{D}$) to adapt complicated resolutions and sea states. Experiments with simulated and real data validate the correctness and effectiveness of the results. The PDOF detector achieves the best performance in most virtual and real-world environments, especially in cases where the target statistics and clutter are not Wishart-distributed.

Ship detection is an important task in civil or military applications and we can use polarimetric synthetic aperture radar (PolSAR). Many polarimetric detectors were proposed and achieved good performances in particular environments, such as optimal polarimetric detector (OPD), polarimetric whitening filter (PWF), polarimetric notch filter (PNF), polarimetric detection optimization filter (PDOF) and diagonal loading detector (DLD) etc. Up to know, the analytical links among different polarimetric detectors have not been found. In this work, the above polarimetric detectors are unified in mathematical forms and a general framework of polarimetric detectors based on quadratic optimization is presented. The mathematical forms are summarized as a trace of two matrices’ product. One is a detection transformation matrix and the other is the polarimetric covariance matrix of the pixel to be detected. We find that all these polarimetric detectors can be regarded as the optimization of such detection matrix, which is the key point of the general framework, and the difficulty turns to be a linear inseparable problem. Pocket Perceptron Linear Algorithm (PPLA) is used to solve the linear inseparable problem. In the case of low resolution, target detection is almost an indivisible problem, and multilayer perceptron (MLP) cannot provide better detection results than PPLA. In the case of high resolution, target detection becomes a nonlinear separable problem, and MLP is gradually superior to PPLA. Additionally, the optimal weights of the recent DLD are obtained to compare with other detectors in the general framework and the DLD is developed to a more general case (GDLD). The experiments validate the general framework of polarimetric detectors. Different detectors in the general framework are utilized and compared in both simulated and measured PolSAR data. The results show the optimal solution in the general framework can always reach the best performance, and the GDLD is the closest one to the optimal detector of the general framework.

Polarimetric synthetic aperture radar (PolSAR) is widely used in remote sensing and has important applications in the detection of ships. Although many polarimetric detectors have been proposed, they are not well combined. Recently, a polarimetric detection optimization filter (PDOF) was proposed, which performs well in most environments. In this study, a novel subspace form of the PDOF [strict PDOF (SPDOF)] was further developed based on the Cauchy inequality and matrix decomposition theories, enhancing detection performance. Furthermore, a simple method to determine the optimal dimension of the subspace detector based on the trace ratio form was proposed by calculating the area under the receiver operating characteristic (ROC) curve, reaching the best detection performance among the subspaces of the detector. Moreover, to combine different subspace detectors, a modified linear discriminant analysis was proposed and developed for the diagonal loading detector (DLD) based on polarimetric subspaces. The experimental results demonstrate the superiority of these joint polarimetric subspace detectors. Most importantly, DLD solves for previous limitations due to the complex clutter background and achieves a performance comparable to that of the Wishart (Gaussian) distribution, particularly in the low target-to-clutter ratio (TCR) case.

The detection of small ships in polarimetric synthetic aperture radar (PolSAR) images is still a topic for further investigation. Recently, patch detection techniques, such as superpixel-level detection, have stimulated wide interest because they can use the information contained in similarities among neighboring pixels. In this article, we propose a novel neighborhood polarimetric covariance matrix (NPCM) to detect the small ships in PolSAR images, leading to a significant improvement in the separability between ship targets and sea clutter. The NPCM utilizes the spatial correlation between neighborhood pixels and maps the representation for a given pixel into a high-dimensional covariance matrix by embedding spatial and polarization information. Using the NPCM formalism, we apply a standard whitening filter, similar to the polarimetric whitening filter (PWF). We show how the inclusion of neighborhood information improves the performance compared with the traditional polarimetric covariance matrix. However, this is at the expense of a higher computation cost. The theory is validated via the simulated and measured data under different sea states and using different radar platforms.

Constant false alarm rate (CFAR) algorithms using a local training window are widely used for ship detection with synthetic aperture radar (SAR) imagery. However, when the density of the targets is high, such as in busy shipping lines and crowded harbors, the background statistics may be contaminated by the presence of nearby targets in the training window. Recently, a robust CFAR detector based on truncated statistics (TS) was proposed. However, the truncation of data in the format of polarimetric covariance matrices is much more complicated with respect to the truncation of intensity (single polarization) data. In this article, a polarimetric whitening filter TS CFAR (PWF-TS-CFAR) is proposed to estimate the background parameters accurately in the contaminated sea clutter for PolSAR imagery. The CFAR detector uses a polarimetric whitening filter (PWF) to turn the multidimensional problem to a 1-D case. It uses truncation to exclude possible statistically interfering outliers and uses TS to model the remaining background samples. The algorithm does not require prior knowledge of the interfering targets, and it is performed iteratively and adaptively to derive better estimates of the polarimetric covariance matrix (although this is computationally expensive). The PWF-TS-CFAR detector provides accurate background clutter modeling, a stable false alarm property, and improves the detection performance in high-target-density situations. RadarSat2 data are used to verify our derivations, and the results are in line with the theory.

Polarimetric whitening filter (PWF) can be used to filter polarimetric synthetic aperture radar (PolSAR) images to improve the contrast between ships and sea clutter background. For this reason, the output of the filter can be used to detect ships. This paper deals with the setting of the threshold over PolSAR images filtered by the PWF. Two parameter-constant false alarm rate (2P-CFAR) is a common detection method used on whitened polarimetric images. It assumes that the probability density function (PDF) of the filtered image intensity is characterized by a log-normal distribution. However, this assumption does not always hold. In this paper, we propose a systemic analytical framework for CFAR algorithms based on PWF or multi-look PWF (MPWF). The framework covers the entire log-cumulants space in terms of the textural distributions in the product model, including the constant, gamma, inverse gamma, Fisher, beta, inverse beta, and generalized gamma distributions ($\text{G}\Gamma $Ds). We derive the analytical forms of the PDF for each of the textural distributions and the probability of false alarm (PFA). Finally, the threshold is derived by fixing the false alarm rate (FAR). Experimental results using both the simulated and real data demonstrate that the derived expressions and CFAR algorithms are valid and robust.

Ship detection is one important task in radar remote sensing. Moreover, Polarimetry shows a valuable contribution to discriminate between targets and clutter. The performance of most polarimetric detectors depends on two important factors: target clutter ratio (TCR) and speckles (or standard deviation to mean ratio of clutter background). The polarimetric matched filter (PMF) is just to maximize the TCR, while the polarimetric whitening filter (PWF) only takes the speckle reduction into consideration. In this paper, the optimal polarimetric detection filter (OPDF) is put forward, which considers maximizing the ratio of TCR to speckle. The approximate expression of the probability density function (PDF) of the OPDF is derived in closed form, so are the probability of false alarm (PFA) and the probability of detection (PD) in Wishart distribution assumption. The threshold of the OPDF detection can be easily obtained in closed form or via the bisection method. Experiments via simulated data validate the correctness of our results. The OPDF detector gives the best performance in most environments, especially in low PFA case and in the case where the statistics of targets is not the ideal Wishart distribution.

The geometrical perturbation-polarimetric notch filter has been recently developed to implement ship detection in polarimetric synthetic aperture radar images. We mathematically prove that the constant false alarm rate detection performance of the filter is independent of the reduction ratio parameter under any a specific distribution of the filter.

A novel estimator for the shape parameter of the generalized Gaussian distribution (GGD) is proposed based on the method of logarithmic cumulants. First, the expression of the log-cumulant for the GGD is theoretically derived. As a result, a simple equation for the estimation of the shape parameter is obtained. The processing procedure of the new estimator for practical applications is also provided. Numerical experiments are used to verify the excellent estimation performances of the proposed estimator for both large and small samples. Moreover, experiments using a measured CARABAS-II data set also validate the superiority of the proposed estimator.

Compact polarimetric data exploitation, especially in hybrid-polarimetric (HP) mode, is currently attracting increasing interest due to the new generation of synthetic aperture radar (SAR) systems. Recently, it has been demonstrated that the notch filter is useful for ship detection in either full- or dual-polarization (DP)-mode SAR images. In this paper, the notch filter investigation is further extended to HP SAR architecture for ship detection on the ocean surface. First, a version of the notch filter that is suitable for HP SAR is proposed based on the definition of the corresponding feature partial scattering vector from the covariance matrix of the HP SAR. Subsequently, a novel model characterizing the statistics of the notch distance of sea clutter in the HP mode is developed. Based on the statistical model, the threshold of constant false-alarm rate (CFAR) ship detection is theoretically and analytically derived, which allows the automatic and adaptive implementation for ship detection in varying sea backgrounds in practical applications. Experiments on the HP SAR data emulated from full-polarization L-band Aerospace Exploration Agency Advanced Land Observation Satellite Phased-Array type L-band SAR and C-band RADARSAT-2 SAR measurements validate not only the soundness of the proposed CFAR detection but also the high accuracy of the presented model in fitting HP SAR data. Furthermore, the notch filter and its CFAR realization provide the same benchmark for the comparison of the detectability of HP and conventional linear DP SAR data. Preliminary findings suggest that the detection performance of HP SAR is superior to that of DP SAR in ship observation. Therefore, the proposed CFAR method based on the notch filter provides a promising technique for the detection of ships using HP SAR data.

Based on its advantages, compact polarimetric (CP) synthetic aperture radar (SAR) is considered to be a good option for earth observations. This paper proposes an adaptive method of ship detection in CP SAR images operating in the hybrid-polarimetric (HP) mode. First, according to the analysis of scattering between ships and sea background, a novel decomposition approach, named power-entropy (PE) decomposition, is developed. Based on this approach, two components of the scattering power, the high-entropy scattering amplitude (HESA) and low-entropy scattering amplitude, are separated. We demonstrate that the HESA component is an effective physical quantity indicating the difference between the ship and its background and hence can potentially be used for ship detection using HP SAR data. The generalized Gamma distribution (GΓD) is found suitable for the characterization of HESA statistics of sea clutter with a wide range of homogeneity. As a result, the adaptive constant false-alarm rate (CFAR) technique based on the HESA detector is proposed. Experiments performed using HP measurements emulated from L-band ALOS-PALSAR and C-band RADARSAT-2 full polarimetric data validate the soundness and superiority of the proposed CFAR method based on the HESA detector. Both the theoretical proof and experimental results show that the HESA improves the ship-sea contrast (or the signal-clutter ratio) more than popular detectors, such as the entropy and SPAN. Moreover, the GΓD is a versatile model for the description of the statistical behavior of both the HESA and comparable detectors.

Accurate knowledge of the statistical properties of synthetic aperture radar (SAR) amplitude data is essential for the processing and interpretation of SAR images. During the last few decades, parametric models have been extensively investigated for use in characterizing the statistics of SAR amplitude data. Among them, the generalized gamma distribution ($\text{G}\Gamma \text{D}$) has been widely applied in many fields of SAR image processing, as it has been demonstrated to be appropriate for describing the statistical behaviors of both SAR land and sea clutter. However, the$\text{G}\Gamma \text{D}$is not effective when the third-order sample log-cumulant of SAR data is close to zero as it cannot match the distribution of actual data. In this case, the recently proposed${\mathcal{ G}}_{\textrm {AO}} $model shows good performance for fitting the distribution of the actual SAR data. In this paper, we theoretically derive the analytical conditions of the${\mathcal{ G}}_{\textrm {AO}} $model applicability. Based on the derivations, a scheme describing the statistical behaviors of SAR pixel amplitudes is given by combining the$\text{G}\Gamma \text{D}$and the${\mathcal{ G}}_{\textrm {AO}} $model. Experiments performed on C-band RADARSAT-2 and L-band ALOS-PALSAR SAR data verify the effectiveness of the proposed scheme. The usefulness of the${\mathcal{ G}}_{\textrm {AO}} $model for extremely heterogeneous land clutter, for example, from urban regions, is also demonstrated.

In the detection applications of synthetic aperture radar (SAR) data, a crucial problem is developing precise models for clutter statistics. Generalized gamma distribution (GΓD) has been widely applied in many fields of signal processing, and it has been demonstrated to be an appropriate model for describing the statistical behaviors of SAR sea clutter, wherein parameter estimation is a key issue for determining the practical application of GΓD. Work that contains three major aspects is performed in this paper. First, an approximate estimator for GΓD parameters based on the well-known “method-of-log-cumulants” is derived; a theoretical comparison between the approximate estimator and other known estimators is also presented. Second, based on this estimator, a scheme of parameter estimation is further given by comprehensively considering estimation precision, speed, and applicable conditions. The simulation results show that the presented scheme is fast and effective. Third, we assess the fitting performance of GΓD and the proposed scheme using real SAR sea clutter data, and compare the model with generalized-K distribution. The experiments on single-look complex and multilook processing L-band ALOS-PALSAR and C-band RADARSAT-2 SAR data verify the effectiveness of the proposed scheme of GΓD parameter estimation. Moreover, several examples of ship detection in real SAR images testify to the usefulness of the proposed scheme in practical applications.

Synthetic aperture radar (SAR) in along-track interferometry (ATI) mode has been extensively applied in velocity measurements of ocean currents and ship detection. A notch filter was recently proposed and was demonstrated to be a promising tool for ship detection exploiting quad- or dual-polarization SAR information. In this paper, the investigation of the notch filter is extended to the dual-channel ATI-SAR mechanism for ship detection on the ocean surface. First, a theoretical proof that the interferometric magnitude performs better in ship detection than single-channel amplitude/intensity is given based on the signal-clutter-ratio improvement, which validates the advantages of interferometric SAR against conventional single-channel SAR for the purpose of ship observations on the ocean surface. Second, based on the proof, the version of the notch filter that is suitable for the ATI-SAR, that is, the interferometric notch filter (INF), is proposed. We then analyze the statistical models of the INF distance, which can facilitate the adaptive realization of INF for ship detection in ATI-SAR. Finally, the experimental results for dual-channel ATI-SAR data measured by NASA/JPL L-band AIRSAR verify the accuracy of the theoretical analysis and effectiveness of the INF.

Synthetic aperture radar (SAR) ship detection is an important research topic in the field of maritime applications. The geometrical perturbation-polarimetric notch filter (GP-PNF) was recently proposed to be a promising tool and its usefulness in exploiting polarimetric SAR information for ship detection was demonstrated. The work in this paper is devoted to developing a statistical model of the filter in nonhomogeneous sea clutter to achieve constant false alarm rate (CFAR) detection based on the model. First, within the framework of a multiplicative model, the reciprocal of the gamma distribution is used to describe the texture component of sea clutter in nonhomogeneous background. As a result, a statistical model of the GP-PNF is analytically derived and found suitable for sea clutter scenes with a wide range of homogeneity. Second, we theoretically demonstrate that CFAR detection using GP-PNF is unrelated to the parameter in the original GP-PNF. Therefore, a simplified version of the GP-PNF is given. Third, the CFAR threshold of the simplified filter is mathematically derived. Experiments performed on measured L-band ALOS-PALSAR and C-band RADARSAT-2 SAR data verify the good performance of the developed statistical model and demonstrate the usefulness of the CFAR detection based on the simplified filter.

Road extraction in SAR (Synthetic Aperture Radar) images is a difficult task because of speckle noise and various interferences. PolSAR (Polarimetric SAR) measures road's reflectivity in four polarizations and provides more information of road, which potentially indicates higher extraction performance than that in single-polarization SAR cases. However, the implementation of the traditional MRF (Markov Random Field) based road extraction method needs excessive computation operations, which is relatively time consuming and difficult for application. This letter presents a GPU-accelerated road extraction method in PolSAR images with five stages. Firstly, the total power is calculated to form the span image. Then, pixels of the similar scattering property are included in the refined Lee filtering process. Thirdly, the edges are detected by the bi-windows edge detection algorithm. Fourthly, line pre-processing and graph construction are done to decrease the false line elements and form the line graph. Finally, the MRF is used to extract road line elements. The experiments have proved that such a GPU-accelerated approach can improve the computing efficiency, and the effectiveness is demonstrated by using the airborne PolSAR image acquired over the Oberpfaffenhofen area in Germany.

Robust feature tracking is a basic requisite for indoor navigation. The Scale Invariant Features Transform (SIFT) features are invariant to image translation, scaling, rotation, and partially invariant to illumination changes. Therefore, it is widely used for image matching based indoor navigation. However, the implementation of the traditional SIFT algorithm needs excessive computation operations, which is relatively time consuming and difficult for indoor navigation like real-time application. This article proposes a SIFT acceleration strategy based on graphics processing unit (GPU). It is proposed as the following four steps. Firstly, Gaussian pyramid is divided into different blocks. And in each GPU block, DoG (Difference of Gaussian) scale-space is computed. Secondly, local keypoints detection is done in GPU, and each keypoint is processed in one block to calculate gradient orientation and magnitude. Thirdly, the keypoint descriptor is formulated as vectors, and each vector is computed in one GPU block. Finally, GPU-accelerated SIFT is introduced into the indoor navigation system to address viewpoint changes. Our main contribution of this article is using an optimized GPU accelerated scheme for real-time indoor navigation. The experiments have proved that such approach can improve the computing efficiency and reduce the chances of mismatches.

The product of multilook amplitudes (PMA) detector has been used to detect ships in high-resolution dual-polarization synthetic-aperture-radar images. However, the adaptive constant false-alarm rate (CFAR) technique of the PMA detector is desirable for practical applications, wherein a crucial problem is to find an appropriate model to describe the PMA statistics for varied sea surfaces. First, we consider a new probability density function to characterize the PMA statistics of homogeneous sea surfaces. Second, by using the new density and multiplicative model, the PMA detector's statistical model for nonhomogeneous sea surfaces is specified and demonstrated to be the G0distribution. Then, a theoretical analysis of the relationship between the performance of the standard CFAR detection and the parameters in the G0distribution is conducted. Experiments performed on the measured RADARSAT-2 and NASA/JPL AIRSAR images verify the effectiveness and appropriateness of the G0model for describing the statistical behavior of the PMA of sea clutter, as well as the usefulness of the model for practical ship-detection applications.

Synthetic aperture radar (SAR), in along-track interferometry mode, is extensively used in sensing oceanic surface. Detecting moving ships is becoming an increasingly important requirement in global monitoring of environment and maritime security. To realize this requirement, we first constructed a metric for detecting moving ships in SAR images by combining sea interferogram's magnitude and phase (SIMP). Second, the performance of the SIMP metric was evaluated through simulations and investigation of several key issues, such as azimuth ambiguity, minimum detectable velocity, and detection performance. The evaluation results establish the effectiveness of the proposed SIMP metric.

A new powerful constant false-alarm rate (CFAR) method is proposed for detecting moving ships in along-track interferometric synthetic aperture radar (ATI-SAR) images, based on the sea interferogram's magnitude and phase (SIMP) metric, presented in Part I of this paper. First, within the structure of the product model, a statistical model (hereafter simply termed Nq) for the square root of SIMP metrics of sea clutter was developed by assuming the radar cross-sectional (RCS) components of the return follow a commonly used gamma distribution. Moreover, the parameter estimators of the presented model were analytically derived by applying the well-known “method of log cumulants” (MoLC). Finally, the CFAR threshold of moving ship detection, using the proposed model was given analytically. The experimental results of measured ATI-SAR data by NASA/JPL's AirSAR show that the proposed method gives good performance.

In the above named paper [ibid., vol. 49, no. 1, pp. 277-294, Jan. 2011], there is an error in equation (13). The corrected equation is given here.

In this letter, a novel constant false alarm rate (CFAR) detection algorithm for high-resolution synthetic aperture radar (SAR) images is proposed with the generalized gamma distribution (GΓD) modeling the background. At first, the method of log-cumulants is introduced for estimating the parameters of the GΓD. In addition, a closed-form expression for the detection threshold of the CFAR algorithm is derived, which refers to the inverse incomplete gamma function. Finally, comparing with the algorithms using the Weibull,KA, andGA0distributions for background, the advantages of the proposed algorithm, including maintaining the false alarm rate and the efficiency, are validated with an actual high-resolution SAR image.

By analyzing the backscattering difference between moving targets and stationary clutter contained in the interferogram, a constant false alarm rate (CFAR) detecting method based on a joint metric of synthetic aperture radar (SAR) interferogram's magnitude and phase (IMP) for ground moving targets has been proposed in this paper. First, utilizing the exclusive mapping relationship between interferometric phase and the module of corresponding vector, and combining interferometric magnitude, a new joint metric, called IMP metric, has been constructed. Second, under the frame of multiplicative model, based on the complex Wishart-distribution, the IMP metric's statistical model, simply denoted asS0distribution, has been derived. Meanwhile, the parametric estimators of theS0distribution have also been presented based on the Mellin transform. Finally, the CFAR threshold of ground moving target detection using theS0distribution has been given analytically. The experiments performed on measured dual-SAR images not only show the effectiveness of the IMP metric's statistical models and the parametric estimators, but also prove the good performance of the proposed CFAR detecting method.

This letter proposes a Parzen-window-kernel-based algorithm for ship detection in synthetic aperture radar (SAR) images. First, the data-driving kernel functions of Parzen window are utilized to approximate the histogram of real SAR image, in order to complete the accurate modeling of SAR images. Then, a threshold of global constant false alarm rate is given theoretically, and the numerical solution of the threshold is also derived. The experimental results of the real data of typical targets demonstrate that the algorithm presented is effective.

To design a highly automatic and practical method for target discrimination in synthetic aperture radar images, we propose in this paper an improved scheme consisting of the framework and algorithms for target discrimination. Our main contribution in this scheme comprises four aspects. First, an integrative frame sequentially combining the algorithm based on feature extraction and the knowledge of target group has been presented. Second, three new features for target discrimination have been introduced. Third, a genetic algorithm-based feature-selection algorithm has been presented. The results show that this algorithm can evaluate the goodness-of-feature better. Finally, to improve the accuracy of the discriminator, we have designed a weighted quadratic distance discriminator, which has been observed to improve the performance of target discrimination. We have analyzed the performance of the proposed scheme comprehensively and specifically using some measured data, and carried out comparisons of the existing algorithms. The results show that the proposed scheme could improve the application ability in target discrimination.

The Lagrange multiplier method is one of the basic optimization procedures to find the optimum polarizations for the incoherent scattering case. This letter proves for the first time that a fixed relationship exists between the optimum polarization and the Lagrange multiplier. Then, an optimization procedure is proposed to simplify the computational complexity of the Lagrange multiplier method. To speed up the convergence of the proposed procedure, the minimum search intervals are discussed and given theoretically. A numerical example is shown to demonstrate the effectiveness of the proposed procedure.

An adaptive and fast constant false alarm rate (CFAR) algorithm based on automatic censoring (AC) is proposed for target detection in high-resolution synthetic aperture radar (SAR) images. First, an adaptive global threshold is selected to obtain an index matrix which labels whether each pixel of the image is a potential target pixel or not. Second, by using the index matrix, the clutter environment can be determined adaptively to prescreen the clutter pixels in the sliding window used for detecting. The$G^{0}$distribution, which can model multilook SAR images within an extensive range of degree of homogeneity, is adopted as the statistical model of clutter in this paper. With the introduction of AC, the proposed algorithm gains good CFAR detection performance for homogeneous regions, clutter edge, and multitarget situations. Meanwhile, the corresponding fast algorithm greatly reduces the computational load. Finally, target clustering is implemented to obtain more accurate target regions. According to the theoretical performance analysis and the experiment results of typical real SAR images, the proposed algorithm is shown to be of good performance and strong practicability.