Feature selection and weighted SVM classifier-based ship detection in PolSAR imagery

Abstract

Target decomposition is an important method for ship detection in polarimetric synthetic aperture radar (SAR) imagery. Parameters such as the polarization entropy and alpha angle deduced from the coherency matrix eigenvalue decomposition capture the differences between the target and background from different views separately. However, under the conditions of a relatively high resolution and a rough sea, the contrast between ship and sea reduces in the aforementioned space. Based on the analyses of target decomposition theory and the target's scattering mechanism, multi-polarization parameters can be used to characterize different scattering behaviours of the ship target and sea clutter. Moreover, each parameter has its own diverse significance in the practical detection problem. This article proposes a feature selection and weighted support vector machine (FSWSVM) classifier-based algorithm to detect ships in polarimetric SAR (PolSAR) imagery. First, the method constructs a feature vector that consists of multi-polarization parameters. Then, different polarization parameters are refined and weighted according to their significance in the support vector machine (SVM) classifier. Finally, ships are classified from the sea background and other false alarms by the classifier. The validation results on National Aeronautics and Space Administration/Jet Propulsion Laboratory (NASA/JPL) airborne synthetic aperture radar (AIRSAR) and Radarsat-2 quad polarimetric data illustrate that the method detects ship targets more precisely and reduces false alarms effectively.

Fully Convolutional Networks and a Manifold Graph Embedding-Based Algorithm for PolSAR Image Classification

“…From the perspective of prior information, existing algorithms can be divided into two categories: supervised and unsupervised methods. Wishart classifiers [8], SVM [9], sparse compressed sensors [10] and ensemble learning [11] are typical examples of supervised classification methods. Due to the lack of PolSAR imagery, unsupervised classification methods have been favored by many scholars.…”

SVM-Based Incremental Learning Algorithm for Large-Scale Data Stream in Cloud Computing

“…The LS-SVM worked on Gaussian noise while the weighted LS-SVM worked on the outliers and the non-Gaussian noise. Xing [22] presented a feature selection and weighted support vector machine (FSWSVM) classifier-based algorithm to detect ships using polarimetric SAR imagery. For the online classification of data streams with imbalanced class distribution, Zhu [23] gave an incremental Linear Proximal Support Vector Machine, named the LPSVM, also called the DCIL-IncLPSVM, which provided robust learning performance in the case of class imbalance.…”

Optical resolution requirements for effective deep learning-based ship recognition

“…In addition, the AIS may not be turned on or may not be functioning appropriately, which can cause targets to go undetected. Various imaging sensors can be applied to detect ships, e.g., synthetic aperture radar (SAR) [1][2][3][4][5] , infrared (IR) [6] sensors, and visible (VIS) [7][8][9] sensors.…”

“…This method detects the target size, texture, and other information to achieve accurate classification of targets. In addition, Xing et al (2013) [2] proposed a method that applied a weighted support vector machine (SVM) to SAR image classification and achieved high classification effect, and Leng et al (2015) [3] employed an image grayscale feature method to extract ship targets from SAR images, and they constructed a constant false alarm target detector via the mathematical statistics of the distribution difference between the target and background. Based on the deep learning network for sea and land segmentation (SLS-CNN), Liu et al (2017) [4] used spectral residual significance heat maps and corner probability distributions for sea and land segmentation, and they detected target ships using the SLS-CNN.…”