8 March 2024 in IEEE Transactions on Aerospace and Electronic Systems
High-resolution range profile (HRRP) is increasingly employed in radar target recognition under intricate ground scenarios. Such scenarios demand recognizing the specific type of a target from a wide range of categories, a task known as fine-grained target recognition (FGTR), which involves numerous and potentially unbalanced categories. To tackle this, we propose a joint semantic-data guided hierarchical classification (SDHC) framework. It consists of a set of local classifiers organized in a tree hierarchy based on the joint semantic-data relationship. It allows the complex FGTR task to be simplified into multiple small-scale sub-tasks. Specifically, the proposed SDHC method focuses on tree hierarchy construction and local classifier training. We design the tree hierarchy based on a joint semantic-data similarity measure, which quantifies the data similarity between categories and incorporates semantic knowledge constraints. Following this, we deploy hierarchical feature selection on a multi-dimensional feature set, considering the contribution of features in each local classifier. Experimental results on measured data verify the effectiveness of the proposed method. Moreover, analysis results demonstrate the superiority of the hierarchical approach over flat methods.
27 November 2025 in IEEE Transactions on Aerospace and Electronic Systems
High-resolution range profile (HRRP) has emerged as a promising approach in real-time radar ground target recognition. To address the real-world operational requirements, it is demanding to develop open set recognition (OSR) that identifies both known and unknown classes of targets. In this paper, we propose a neural collapse-guided hierarchical prototypical learning (NHPL) method for HRRP OSR. The key of prototypical learning is constructing class-specific prototypes in the feature space as the representatives. However, this is challenging in HRRP recognition due to the intrinsic aspect-sensitivity, which results in variation of feature distribution and deviation of estimated prototypes. To tackle the above issues, we employ a hierarchical classification strategy which decomposes the OSR task into a multiple level recognition framework according to the hierarchical semantic taxonomy of classes. Within each level, discriminate features can be more effectively learned, thereby constraining the variation. Furthermore, inspired by neural collapse theory, we estimate class prototypes using the weight vectors of the trained network's final layer. This approach aligns prototypes closely with features of known classes while staying distant from features of unknown classes, thereby reducing prototype deviation. Extensive experiments on measured HRRPs demonstrate that the proposed method outperforms existing methods in terms of accuracy and robustness for recognizing unknown classes.