EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Analysis of the Matchability of Reference Imagery for Aircraft Based on Regional Scene Perception

Xin Li, Guo Zhang, Hao Cui, Jinhao Ma, Wei Wang

Remote Sensing · 04.09.2023 · 15 (17) · 4353

DOI: 10.3390/rs15174353 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 04.09.2023. Verfasst von Xin Li und 4 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fundamental Research Funds for the Central Universities, China, Foundation Strengthening Fund Project, China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Scene matching plays a vital role in the visual positioning of aircraft. The position and orientation of aircraft can be determined by comparing acquired real-time imagery with reference imagery. To enhance precise scene matching during flight, it is imperative to conduct a comprehensive analysis of the reference imagery’s matchability beforehand. Conventional approaches to image matchability analysis rely heavily on features that are manually designed. However, these features are inadequate in terms of comprehensiveness, efficiency, and taking into account the scene matching process, ultimately leading to unsatisfactory results. This paper innovatively proposes a core approach to quantifying matchability by utilizing scene information from imagery. The first proposal for generating image matchability samples through a simulation of the matching process has been developed. The RSPNet network architecture is designed to effectively leverage regional scene perception in order to accurately predict the matchability of reference imagery. This network comprises two core modules: saliency analysis and uniqueness analysis. The attention mechanism employed by saliency analysis module extracts features at different levels and scales, guaranteeing an accurate and meticulous quantification of image saliency. The uniqueness analysis module quantifies image uniqueness by comparing neighborhood scene features. The proposed method is compared with traditional and deep learning methods for experiments based on simulated datasets, respectively. The results demonstrate that RSPNet exhibits significant advantages in terms of accuracy and reliability.

Förderung

  • Fundamental Research Funds for the Central Universities, China
  • Foundation Strengthening Fund Project, China

Zitieren

APA

Li, X., Zhang, G., Cui, H., Ma, J., & Wang, W. (2023). Analysis of the Matchability of Reference Imagery for Aircraft Based on Regional Scene Perception. Remote Sensing, 15(17), 4353. https://doi.org/10.3390/rs15174353

Vancouver

Li X, Zhang G, Cui H, Ma J, Wang W. Analysis of the Matchability of Reference Imagery for Aircraft Based on Regional Scene Perception. Remote Sensing. 2023;15(17):4353. doi:10.3390/rs15174353

BibTeX
@article{Li2023,
  author = {Li, Xin and Zhang, Guo and Cui, Hao and Ma, Jinhao and Wang, Wei},
  title = {Analysis of the Matchability of Reference Imagery for Aircraft Based on Regional Scene Perception},
  journal = {Remote Sensing},
  year = {2023},
  volume = {15},
  number = {17},
  pages = {4353},
  doi = {10.3390/rs15174353}
}
Kostenlos starten →