EUVIMEDEUVIMED entdecken
Crossref

MKNN-OptConcave Detection for Small Targets in Sea Clutter

Yanling Shi, Chenlin Song, Leiyao Liao, Jianguo Yao

IEEE Transactions on Aerospace and Electronic Systems · 2026 · 62 · 7345-7362

DOI: 10.1109/taes.2026.3673294 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Aerospace and Electronic Systems, 2026. Verfasst von Yanling Shi und 3 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

To address the limitations in performance and real-time capability of traditional concave hull algorithms for small target detection in sea clutter backgrounds, we propose a joint detection algorithm based on magnetism K-nearest neighbors (MKNN) and optimized concave hull (OptConcave). The MKNN employs a dynamic force propagation mechanism to iteratively expand clusters of clutter feature points in dense regions, effectively growing the point set to meet false alarm rate (FAR) requirements. The OptConcave begins with a standard convex hull, using clutter feature points as the vertices of tetrahedra. Through Delaunay triangulation, the convex hull is partitioned into nonoverlapping tetrahedral meshes, and external tetrahedra are iteratively removed to construct a concave hull that geometrically encloses clutter points with greater precision and compactness. The proposed MKNN-OptConcave detector generates a smooth, sealed concave surface that avoids these issues, such as self-intersection, clutter leakage, and cavity phenomenon similar to a swimming ring. This surface ensures structural closure, waterproofing, and topological connectivity. Evaluated on the IPIX dataset, the MKNN-OptConcave detector achieves an average detection probability of 93% at an FAR of 0.001. Compared to the comparison algorithms, the average detection probability increased by up to 3.2% in the best-performing case. Furthermore, training time is reduced by 70.7% and detection time by 95.4%, making it highly suitable for real-time applications.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Shi, Y., Song, C., Liao, L., & Yao, J. (2026). MKNN-OptConcave Detection for Small Targets in Sea Clutter. IEEE Transactions on Aerospace and Electronic Systems, 62, 7345-7362. https://doi.org/10.1109/taes.2026.3673294

Vancouver

Shi Y, Song C, Liao L, Yao J. MKNN-OptConcave Detection for Small Targets in Sea Clutter. IEEE Transactions on Aerospace and Electronic Systems. 2026;62:7345-7362. doi:10.1109/taes.2026.3673294

BibTeX
@article{Shi2026,
  author = {Shi, Yanling and Song, Chenlin and Liao, Leiyao and Yao, Jianguo},
  title = {MKNN-OptConcave Detection for Small Targets in Sea Clutter},
  journal = {IEEE Transactions on Aerospace and Electronic Systems},
  year = {2026},
  volume = {62},
  pages = {7345--7362},
  doi = {10.1109/taes.2026.3673294}
}
Kostenlos starten →