Ship detection and classification pose critical challenges in maritime surveillance and monitoring. On the one hand, they find applications in civilian domains, like fishery management, illegal immigrant monitoring, and marine rescue. On the other hand, ships are important targets in maritime combat. Ship detection and classification help to obtain timely maritime military intelligence that holds immense importance in seizing control of the maritime battlefield and ensuring the success of military operations [1], [2]. Remote sensing technology, facilitated by sensors mounted on platforms like satellites and airplanes, enables long-distance noncontact imaging of ground targets. It offers high timeliness, wide coverage, and low-cost accessibility and is not constrained by geographic conditions. Consequently, it has emerged as the primary method for monitoring vessels [4], [5].

A major bottleneck in limiting the application of the existing methods of ship classification in synthetic aperture radar (SAR) images is the inadequate amount of labeled data available for training a classifier. However, generating ground truth involves expensive and time-consuming ground campaigns or is costly, since a high number of SAR image acquisition will be necessary. In contrast, an automatic identification system (AIS), which is an automatic tracking system used for monitoring maritime ships, can provide plenty of labeled ship samples that is relatively easier to be obtained. Inspired by these facts, this letter proposes to improve ship classification in SAR images by transferring AIS knowledge. We propose an improved multiclass adaptive support vector machine, combined with the naive geometric features (NGFs), to achieve transfer learning between the AIS domain and the SAR image domain. The experiments prove that the traditional method can be significantly improved by AIS information transfer, especially when only a few training samples in the SAR domain are available. In addition, it also shows that after feature selection, the performance of the proposed method can be close to that of the state of the art, even if by only using simpler NGFs and few training samples.

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 the amount of information in a moderate-resolution SAR image and the stability of feature extraction, we propose naive geometric features (NGFs) for ship classification. In contrast to the strictly defined geometric features (SGFs), the extraction of NGFs is very simpler and efficient. And more importantly, the NGFs are enough to reveal the essential difference between different types of ships for classification. To fuse various NGFs with different physical properties and discriminability, the multiple kernel learning (MKL) is utilized to learn the combination weights, rather than assigning the same weight to all features as usually applied by the traditional support vector machines (SVMs). The comprehensive experiments validate that: (1) the performance of the proposed NGF-combined MKL outperforms that of NGF-combined SVM by 3.4% and is very close to that obtained by SGF-combined MKL and (2) in terms of classifying ships in a moderate-resolution SAR image, NGFs are more feasible than scattering features.

There are still many challenges to be resolved in the task of ship classification in synthetic aperture radar (SAR) images, such as limited number of labeled samples in SAR domain, large variance in the same subcategory, small variance among different subcategories, etc. Transfer metric learning (TML) has the potential to mitigate those issues in the domain of interest (target domain, TD) by leveraging knowledge/information from other related domains (source domain, SD). In this article, we proposed a novel TML method, termed as geometric transfer metric learning (GTML), which achieves discriminative information preservation (DIP), geometric structure preservation (GSP), and handles the domain shift (DS) simultaneously by integrating pairwise constraints (PC), joint distribution adaptation (JDA), and manifold regularization (MR) into a unified optimization function, aiming to make full use of their complementarity to improve SAR ship classification performance. In practice, we proposed two simple but effective optimization strategies, termed as GTML-A and GTML-R, to construct optimization function. We also proposed two solutions for two typical real-world application scenarios, that is, the task of: 1) zero-labeled sample (ZLS) and 2) scarce-labeled samples (SLS) in SAR domain. The experiments conducted on both tasks show that the proposed GTML outperforms most of state-of-the-art methods. Code is available at https://github.com/sky-Yongjie-Xu/geometric-transfer-metric-learning.