Dynamic Pattern-Coupled Sparse Bayesian Learning for Spaceborne ISAL Reconstruction
IEEE Signal Processing Letters · 2026 · 33 · 4161-4165
DOI: 10.1109/lsp.2026.3742301 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Minghui Wang und 3 weiteren Autor:innen. Die Arbeit nennt 18 Literaturangaben. Gefördert von: Northwestern Polytechnical University.
Abstract (Originalsprache)
Inverse synthetic aperture lidar (ISAL) enables high-resolution imaging of noncooperative space targets, but sparse-aperture observations often cause severe defocusing and sidelobe artifacts. Existing sparse Bayesian methods either neglect spatial dependencies or use predefined coupling strengths, limiting their adaptability to spatially varying scattering structures. This letter proposes a dynamic pattern-coupled sparse Bayesian learning method for spacecraft ISAL reconstruction, in which coupling strengths are adaptively learned from the observations to characterize local dependencies among neighboring scattering coefficients. An auxiliary-variable-based Jensen lower bound is incorporated into a generalized maximum-a-posteriori expectation-maximization framework, yielding closed-form hyperparameter updates without changing the leading-order computational complexity compared with the conventional pattern-coupled algorithm. Experiments on simulated spacecraft data and measured Yak-42 data demonstrate improved reconstruction fidelity and scattering-support preservation under sparse and noisy observations.
Förderung
- Northwestern Polytechnical University
Zitieren
APA
Wang, M., Cheng, C., Wang, D., & Sun, J. (2026). Dynamic Pattern-Coupled Sparse Bayesian Learning for Spaceborne ISAL Reconstruction. IEEE Signal Processing Letters, 33, 4161-4165. https://doi.org/10.1109/lsp.2026.3742301
Vancouver
Wang M, Cheng C, Wang D, Sun J. Dynamic Pattern-Coupled Sparse Bayesian Learning for Spaceborne ISAL Reconstruction. IEEE Signal Processing Letters. 2026;33:4161-4165. doi:10.1109/lsp.2026.3742301
BibTeX
@article{Wang2026,
author = {Wang, Minghui and Cheng, Cheng and Wang, Dong and Sun, Jinqiu},
title = {Dynamic Pattern-Coupled Sparse Bayesian Learning for Spaceborne ISAL Reconstruction},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {4161--4165},
doi = {10.1109/lsp.2026.3742301}
}
EUVIMED