EUVIMEDEUVIMED entdecken
Crossref

Unsupervised multistage registration method for remote sensing images with large geometric deformation

Yan Li, Yingdan Wu, Yang Ming, Zhesheng Cheng, Yu Liu

Journal of Applied Remote Sensing · 12.06.2025 · 19 (02) · 1-17

DOI: 10.1117/1.jrs.19.026508 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Applied Remote Sensing, erschienen am 12.06.2025. Verfasst von Yan Li und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 42 Literaturangaben. Gefördert von: Key Research and Development Program of Hubei Province of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Remote sensing images serve as an important data source for Earth observation. However, the automatic registration of remote sensing images with significant geometric deformations still encounters issues such as low reliability and success rates. To tackle this problem, we propose an unsupervised multistage registration method. Based on the “divide and conquer” strategy, the model consists of four relatively independent transformation parameter estimation modules, namely the translation, rotation, scaling, and shearing modules. They share the same feature extraction model and utilize the Swin transformer to extract global features for parameter regression. The model is trained in an unsupervised way by combining the independent training of each module with the unified training of the whole model directly based on the gray similarity of the matching images. Experimental results show that this method effectively overcomes the impacts of significant geometric deformations and achieves reliable registration for remote image registration tasks. Compared with classical deep learning methods and traditional approaches, it can improve both the success rate and reliability to a certain extent.

Förderung

  • Key Research and Development Program of Hubei Province of China

Zitieren

APA

Li, Y., Wu, Y., Ming, Y., Cheng, Z., & Liu, Y. (2025). Unsupervised multistage registration method for remote sensing images with large geometric deformation. Journal of Applied Remote Sensing, 19(02), 1-17. https://doi.org/10.1117/1.jrs.19.026508

Vancouver

Li Y, Wu Y, Ming Y, Cheng Z, Liu Y. Unsupervised multistage registration method for remote sensing images with large geometric deformation. Journal of Applied Remote Sensing. 2025;19(02):1-17. doi:10.1117/1.jrs.19.026508

BibTeX
@article{Li2025,
  author = {Li, Yan and Wu, Yingdan and Ming, Yang and Cheng, Zhesheng and Liu, Yu},
  title = {Unsupervised multistage registration method for remote sensing images with large geometric deformation},
  journal = {Journal of Applied Remote Sensing},
  year = {2025},
  volume = {19},
  number = {02},
  pages = {1--17},
  doi = {10.1117/1.jrs.19.026508}
}
Kostenlos starten →