Selecting discriminate features and constructing an appropriate classifier are two essential factors for ship classification in a synthetic aperture radar (SAR) image. Unfortunately, these rarely considered together by existing studies. We propose joint feature selection method integrating the strategy into wrapper framework. The sequential forward floating searching algorithm is improved to conduct efficient optimal triplet of feature-scaling-classifier. Comprehensive experiments on data...

A major bottleneck in limiting the application of existing methods ship classification synthetic aperture radar (SAR) images is inadequate amount labeled data available for training a classifier. However, generating ground truth involves expensive and time-consuming campaigns or costly, since high number SAR image acquisition will be necessary. In contrast, an automatic identification system (AIS), which tracking used monitoring maritime ships, can provide plenty samples that relatively...

This paper proposes a new scheme for detecting ship targets in high-resolution (HR) single-channel synthetic aperture radar (SAR) images. By using the proposed spatially enhanced pixel descriptor (SEPD) and modified density-based spatial clustering of application with noise (M-DBSCAN), this can overcome typical challenges detection HR SAR Specifically, SEPD maps representation given an image into high-dimensional feature space by embedding intensity information its neighborhood...

Compared with the high-resolution synthetic aperture radar (SAR) image, a moderate-resolution SAR image can offer wider swath, which is more suitable for maritime ship surveillance. Taking into account amount of information in and stability feature extraction, we propose naive geometric features (NGFs) classification. In contrast to strictly defined (SGFs), extraction NGFs very simpler efficient. And importantly, are enough reveal essential difference between different types ships To fuse...

Improving ship classification performance in synthetic aperture radar (SAR) imagery by the methods based on transfer learning (TL) is a newly emerging research topic and has great potential. The existing studies merely address problem of from single source domain (AIS or ORS) to target homogeneous (HoTL) which requires all domains are represented features with same dimensions. Our work takes step forward attempts more meaningful challenging that transfers knowledge multiple for purpose...

There are still many challenges to be resolved in the task of ship classification synthetic aperture radar (SAR) images, such as limited number labeled samples SAR domain, large variance same subcategory, small among different subcategories, etc. Transfer metric learning (TML) has potential mitigate those issues domain interest (target TD) by leveraging knowledge/information from other related domains (source SD). In this article, we proposed a novel TML method, termed geometric transfer...

Ship surveillance by remote sensing technology has become a valuable tool for protecting marine environments. In recent years, the successful launch of advanced synthetic aperture radar (SAR) sensors that have high resolution and multipolarimetric modes enabled researchers to use SAR imagery not only ship detection but also category recognition. A hierarchical recognition scheme is proposed. The complementary information obtained from used improve both precision accuracy. stage, three-class...

Ship classification in synthetic-aperture radar (SAR) images is of great significance for dealing with various marine matters. Although traditional supervised learning methods have recently achieved dramatic successes, but they are limited by the insufficient labeled training data. This letter presents a novel unsupervised domain adaptation (DA) method, termed as discriminative regularization framework-based transfer (D-ARTL), to address problem case that there no data available at all SAR...

Fine-grained ship classification in synthetic aperture radar (SAR) images is a challenging task, since SAR can only provide limited discriminative information due to the limitation of imaging mechanism. Distance metric learning (DML) methods have ability improve feature representations through preserving supervisory samples. In this article, we proposed novel DML method, termed as distribution shift (DML-ds), which improves original Laplacian regularized by adding an interclass...

Improving ship classification performance in synthetic aperture radar (SAR) imagery by transferring knowledge from the related domain is a newly emerging research topic. Existing methods follow supervised or unsupervised homogeneous transfer learning techniques with certain restrictions on use of features (homogeneous rather than heterogeneous) and data (ignoring excavate potential unlabeled target data), which may hinder further improvements. To address these problems, this letter proposes...

Scattering mechanism (SM) analysis is a promising technique for ship detection and classification in polarimetric SAR (PolSAR) images. In this paper, four-component model-based decomposition method incorporating surface, double-bounce, volume cross-polarized components proposed analyzing the SMs of ships. A novel scattering component capable discriminating between HV power generated by oriented scatterers on ships from that as means to address problem overestimation. stage, taking into...

Abstract. In this paper we introduce a parameter for the retrieval of thickness undeformed first-year sea ice that is specifically adapted to compact polarimetric (CP) synthetic aperture radar (SAR) images. The denoted as "CP ratio". model simulations investigated sensitivity CP ratio dielectric constant, thickness, surface roughness, and incidence angle. From results deduced optimal conditions angles retrieval. C-band SAR data acquired over Labrador Sea in circular transmit linear receive...

This study aims at improving fine-grained ship classification performance under the condition that there is no labeled samples available in SAR domain (target domain) by transferring knowledge from optical remote sensing (ORS) (source which has rich samples. The proposed method improves original deep subdomain adaptation network (DSAN) designing a dual-branch (DBN) embedding attention module to extract more discriminative transferable features, thereby of adaptation. Specifically, we...

In recent years, oil spill surveillance with space-borne synthetic aperture radar (SAR) has received unprecedented attention and been gradually developed into a common technique for maritime environment protection. A typical SAR-based detection process consists of three steps: (1) dark-spot segmentation, (2) feature extraction, (3) look-alike discrimination. As preliminary task in the chain, segmentation is critical fundamental step prior to extraction classification, since its output direct...

Summary Effectively obtaining the location and direction of ship target is an important prerequisite for maritime traffic management marine accident rescue. Thanks to rapid development detection methods based on deep learning, this article proposed a method multiresolution synthetic aperture radar (SAR) images improved region convolution neural network (R‐CNN). According characteristics in SAR images, we make several improvements such as enlarging input, proposal optimization, database...

In this paper, we study a novel problem of classifying covert photos, whose acquisition processes are intentionally concealed from the subjects being photographed. Covert photos often privacy invasive and, if distributed over Internet, can cause serious consequences. Automatic identification such therefore, serves as an important initial step toward further protection operations. The is, however, very challenging due to large semantic similarity between and noncovert enormous diversity in...

Benefiting from deep learning, synthetic aperture radar (SAR) ship detection based on convolutional neural network (CNN) has developed rapidly and corresponding performance is getting better. Nevertheless, most of the existing methods still cannot achieve a good balance between <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">precision</i> xmlns:xlink="http://www.w3.org/1999/xlink">recall</i> in scenes with complex background interferences, or...

Ship classification based on high-resolution synthetic aperture radar (SAR) imagery plays an increasingly important role in various maritime affairs, such as marine transportation management, emergency rescue, pollution prevention and control, security situational awareness, so on. The technology of deep learning, especially convolution neural network (CNN), has shown excellent performance ship SAR images. Nevertheless, it still some limitations real-world applications that need to be taken...

In this paper, we address the problem of ship detection in PolSAR image. We firstly investigate differences scattering mechanism between targets and sea surface based on polarimetric similarity analysis. It is shown that, dominated by odd bounce (denoted as r <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> ), while are both even (r xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) which has been widely accepted, line...