This paper addresses the use of a data analysis tool, known as robust principal component analysis (RPCA), in the context of change detection (CD) in ultrawideband (UWB) very high-frequency (VHF) synthetic aperture radar (SAR) images. The method considers image pairs of the same scene acquired at different time instants. The CD method aims to maximize the probability of detection (PD) and minimize the false alarm rate (FAR). Such aim fits into a multiobjective optimization problem, since maximizing the probability of detection generally implies an increase in the number of false alarms. In that sense, varying the RPCA regularization parameter leads to PD variation with respect to FAR, which is known as receiver operating characteristic (ROC) curve. To evaluate the proposed method, the CARABAS-II data set was considered. The experimental results show that RPCA via principal component pursuit (PCP) can provide a good trade-off between PD and FAR. A comparison between the results obtained with the proposed method and a classical CD algorithm based on the likelihood ratio test provides the pros and cons of the proposed method.

Recently, it was demonstrated that low-frequency wavelength-resolution synthetic aperture radar (SAR) images could be considered to follow an additive mixing model due to their backscatter characteristics. This simplification allows for the use of source separation methods, such as robust principal component analysis (RPCA) via principal component pursuit (PCP), for detecting changes in those images. In this manuscript, a change detection method for wavelength-resolution SAR images based on image stack through RPCA is proposed. The method aims to explore both the temporal and flight heading diversity of a set of wavelength-resolution multitemporal SAR images in order to detect concealed targets in forestry areas. A heuristic based on three rules for better exploring the RPCA results is introduced, and a new configurable parameter for false alarm reduction based on the analysis of image windows is proposed. The method is evaluated using real data obtained from measurements of the ultrawideband (UWB) very high-frequency (VHF) SAR system CARABAS-II. Experiments for stacks of four and seven reference images are conducted, and the use of reference images acquired with different flight headings is explored. The results indicate that a gain in performance can be achieved by using large image stacks containing, at least, one image of each possible flight heading of the data set, which can result in a probability of detection (PD) above 99% for a false alarm rate (FAR) as low as one false alarm per three square kilometers. Furthermore, it is demonstrated that high PD and low FAR can be achieved, also considering images from similar flight headings as reference images.

Robust principal component analysis (RPCA) has been widely used for processing and interpreting high-dimensional data in different applications such as data classification, face recognition, video analytics, and recommendation system design. However, the advancement of multisensor-based technologies and the emergence of large data sets have highlighted the limitations of traditional matrix-based models, which have paved the way for higher-order extensions such as tensor RPCA (TRPCA) techniques. These signal separation techniques can be useful for ground scene estimation (GSE) in synthetic aperture radar (SAR) imagery. GSE estimates the clutter-plus-noise content in the scene, and therefore, change detection (CD) methods can benefit, reducing the number of false alarms (FA). This paper presents two new GSE methods for SAR imagery based on robust PCA techniques. The first proposed method uses the RPCA via Principal Component Pursuit (PCP) to obtain the GSE-RPCA. The second method uses TRPCA via New Tensor Nuclear Norm (TNN) to obtain the GSE-TRPCA. The methodology allows the GSE to be obtained through a generalized regularization parameter. The alternating direction method of multipliers (ADMM) algorithm is utilized to solve both optimization problems. Experimental results are evaluated considering real SAR imagery from two data sets acquired with the CARABAS II and ALOS PALSAR systems, respectively. Additionally, the proposed techniques were evaluated under several input characteristics, e.g., eight-image stacks and image pairs. Both GSE techniques are more robust to outliers and missing data when compared to existing solutions found in the literature. Finally, GSE-TRPCA achieved a minimum-square error performance of 0.0018 for some of the evaluated scenarios.

This article aims at performing maritime target classification in SAR images using machine learning (ML) and deep learning (DL) techniques. In particular, the targets of interest are oil platforms and ships located in the Campos Basin, Brazil. Two convolutional neural networks (CNNs), VGG-16 and VGG-19, were used for attribute extraction. The logistic regression (LR), random forest (RF), support vector machine (SVM), k-nearest neighbours (kNN), decision tree (DT), naive Bayes (NB), neural networks (NET), and AdaBoost (ADBST) schemes were considered for classification. The target classification methods were evaluated using polarimetric images obtained from the C-band synthetic aperture radar (SAR) system Sentinel-1. Classifiers are assessed by the accuracy indicator. The LR, SVM, NET, and stacking results indicate better performance, with accuracy ranging from 84.1% to 85.5%. The Kruskal–Wallis test shows a significant difference with the tested classifier, indicating that some classifiers present different accuracy results. The optimizations provide results with more significant accuracy gains, making them competitive with those shown in the literature. There is no exact combination of methods for SAR image classification that will always guarantee the best accuracy. The optimizations performed in this article were for the specific data set of the Campos Basin, and results may change depending on the data set format and the number of images.

Wavelength-resolution (WR) synthetic aperture radar (SAR) change detection (CD) has been used to detect concealed targets in forestry areas. However, most proposed methods are generally based on matrix or vector analyses and, therefore, do not exploit information embedded in multidimensional data. In this letter, a CD method for WR SAR image stacks based on tensor robust principal component analysis (TRPCA) is proposed. The proposed CD method used the new tensor nuclear norm induced by the definition of the tensor-tensor product to exploit temporal and spatial information contained in the image stack. To assess the performance of the proposed method, we considered SAR images obtained by the very high frequency (VHF) WR CARABAS-II SAR system. Experiments for three different stack sizes show that a significant performance gain can be achieved when large image stacks are considered. The proposed CD method performs better in terms of probability of detection (PD) and false alarm rate (FAR) than the other five CD methods in VHF WR SAR images, including one based on matrix robust principal component analysis (RPCA). In a particular setting, it achieves a PD of 99% and a FAR of 0.028 false alarms per km ² .