SAR Target Recognition via Joint Sparse Representation of Monogenic Components with 2D Canonical Correlation Analysis

Abstract

A synthetic aperture radar (SAR) target recognition approach is developed in this paper by exploiting the multiscale monogenic components, which are extracted from SAR images based on the 2D monogenic signal. The 2D canonical correlation analysis is then employed to analyze the correlations of the same monogenic components at different scales. Afterwards, the three monogenic components, i.e., local amplitude, local phase, and local orientation, at different scales are fused as three feature matrices, respectively. In order to further capture the correlations between different types of monogenic components, the joint sparse representation is used for target classification. Therefore, both the correlations of the same monogenic components at multiple scales and the relatedness among different types of monogenic components can be exploited in the proposed scheme. The real measured SAR images from the moving and stationary target acquisition and recognition dataset are classified to examine the validity of the proposal. Compared with some state-of-the-art SAR target recognition methods, the proposed approach is validated to be superior under both standard operating condition and several usual extended operating conditions according to the experimental results. In comparison with some other methods, which also use monogenic components as the basic features, the superiority of the proposed method demonstrates that it could better make use of the monogenic components to improve the classification performance. INDEX TERMS Synthetic aperture radar (SAR), target recognition, monogenic signal, 2D canonical correlation analysis (2DCCA), joint sparse representation (JSR).

SAR Target Recognition via Joint Sparse and Dense Representation of Monogenic Signal

“…We added random noise to the four-class targets, which were used for conducting the experiment (see Table 1). Specifically, a percentage of pixels from original images was randomly chosen and replaced with independent and identically distributed samples within the image pixel values, which is used in several relevant studies [3,17]. However, the difference is that we randomly corrupted 60% of training samples and testing samples, not all the test samples.…”

“…Zhou et al [16] presented a scale selection model, where the specific monogenic component features are produced before classification. Zhou et al [17] presented the feature fusion of multi-scale monogenic components by 2D canonical correlation analysis for SAR target recognition. These studies prove the ability of the sparse representation of monogenic components in SAR target recognition.…”

Block Sparse Bayesian Learning over Local Dictionary for Robust SAR Target Recognition

“…In addition, some newly proposed manifold learning methods were demonstrated effective for SAR ATR [16][17][18][19]. Image decomposition tools including wavelet [20], monogenic signal [21,22], and empirical mode decomposition (EMD) [23,24] were adopted in SAR ATR with good performance.…”

Using Polarization-Diverse Wave-Coefficients for Aerospace Target Recognition

“…The synthetic aperture radar (SAR) image offers an intuitionistic contour profile of the target, besides, it has the advantages of penetration ability through obstacles and day-night operation. Thus, it makes SAR image a well-known radar target feature [12]- [15]. There is a general consensus among the researchers that no single feature vector is optimal on all occasions.…”