22 September 2023 in IEEE Transactions on Circuits and Systems for Video Technology
This paper focuses on generating Inverse Synthetic Aperture Radar (ISAR) images from optical images, in particular, for orbit space targets. ISAR images are widely applied in space target observation and classification tasks, whereas, limited to the expensive cost of ISAR sample collection, training deep learning-based ISAR image classifiers with insufficient samples and generating ISAR samples from emulation optical images via image translation techniques have attracted increasing attention. Image translation has highlighted significant success and popularity in computer vision, remote sensing and data generation societies. However, most of the existing methods are implemented under the discipline of extracting the explicit pixel-level features and do not perform effectively while entailing translation to domains with specific implicit features, such as ISAR image does. We propose a meta-learning based domain prior to implicit feature modelling and apply it to CycleGAN and UNIT models to realize effective translations between the ISAR and optical domains. Two representative implicit features, ISAR scattering distribution feature from the physical domain and the classification identifying feature from the task domain, are elaborately formulated with explicit modelling in statistic form. A meta-learning based training scheme is introduced to leverage the mutual knowledge of domain priors across different samples, and thus allows few-shot learning capacity with dramatically reduced training samples. Extensive simulations validate that the obtained ISAR images have better visible-authenticity and training-effectiveness than the existing image translation approaches on various synthetic datasets. Source codes are available at.
23 April 2025 in IEEE Geoscience and Remote Sensing Letters
This paper addresses the challenge of generating high-fidelity Inverse Synthetic Aperture Radar (ISAR) images from optical images, particularly for space targets. We propose a framework for the generation of ISAR images incorporating component refinement, which attains high-fidelity ISAR scattering characteristics through the integration of an advanced generation model predicated on semantic segmentation, designated as Semantic-Aware ISAR Generation (SAIG). SAIG renders ISAR images from optical equivalents by learning mutual semantic segmentation maps. Extensive simulations demonstrate its effectiveness and robustness, outperforming state-of-the-art methods by over 8% across key evaluation metrics.
21 October 2026 in SSRN Electronic Journal
With the increasing complexity of radar task scenarios and target characteristics, conventional monostatic High Resolution Range Profile recognition methods struggle to meet the demands for high-performance and robust classification, underscoring the critical need for multistation radar fusion frameworks. However, current multistation HRRP fusion approaches predominantly rely on implicit feature learning from the holistic HRRP series, overlooking the differential contributions of local discriminative segments and globally informative frequency-domain representations that capture physical scattering characteristics. These limitations lead to significant performance degradation under noisy conditions. To address these challenges, this paper proposes a Triple Stream Information Extraction (TSIE) framework that comprehensively exploits time-domain, frequency-domain, and time-frequency-domain characteristics for multistation HRRP recognition. Our framework introduces three dedicated components: (1) a Time Segmentation Encoding module that employs a stochastic Monte Carlo sampling process to adaptively identify and emphasize critical local discriminative segments; (2) a Global Frequency Embedding module that utilizes an energy-guided attention mechanism to adaptively extract globally informative and physically meaningful features from the spectrum; and (3) a Time-Frequency Analysis module based on Continuous Wavelet Transform to capture localized spectral dynamics and transient patterns. The complementary features from these three streams are integrated through a Multi-Contrastive Learning strategy that enforces feature consistency across stations while enhancing inter-class discriminability. Extensive experiments on both simulated and measured HRRP datasets demonstrate that TSIE achieves the best recognition accuracies of 97.57% and 96.21%, respectively, outperforming baseline methods. Crucially, under the challenging noise-injection (SNR=0dB) condition, TSIE maintains robust performance with 81.27% accuracy, demonstrating superior performance over baseline approaches.