Teaching and Research Group:UAC

Journal:IEEE Transactions on Aerospace and Electronic Systems

Key Words:Feature extraction;Radar polarimetry;Synthetic aperture radar;Visualization;Target recognition;Data models;Scattering;Transformers;Data mining;Graph neural networks;Feature coordination;hypergraph neural network (HGNN);multiview;synthetic aperture radar (SAR);target recognition

Abstract:Due to the extensive information in multiview images, multiview synthetic aperture radar (SAR) automatic target recognition (ATR) has attracted much attention. However, most current algorithms ignore the integration of multiview topological features intrinsically related to the characteristics of SAR images. Moreover, as well as not thoroughly exploring the inherent coupling relationship, the multiview feature fusion modules in these algorithms are also prone to cause the serve parameter burden and insufficient generality of the entire ATR model. To tackle these issues, an ATR model called multiview visual and topological feature coordination aggregation is proposed. First, two parallel feature extraction modules are employed to extract multiview visual and topological features independently. Specifically, the topological feature extraction module based on a hypergraph neural network is designed to extract the implicit topological features by aggregating the context refinement features of key points within the SAR images. Subsequently, a parameter-friendly feature coordination aggregation module with visual and topological consistency is introduced, which effectively integrates multiview features to generate a unified representation for classification while enhancing the generality of the entire ATR model. Experimental results on the moving and stationary target recognition and the full aspect stationary targets-vehicle datasets verify the effectiveness of our MVT-CA model, even in scenarios involving severe background noise and target deformation.

Indexed by:Journal article (JA)

Translation or Not:no

Included Journals:SCI