EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Lightweight CNN Based on Axial Depthwise Convolution and Hybrid Attention for Remote Sensing Image Dehazing

Yufeng He, Cuili Li, Xu Li, Tiecheng Bai

Remote Sensing · 31.07.2024 · 16 (15) · 2822

DOI: 10.3390/rs16152822 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 31.07.2024. Verfasst von Yufeng He und 3 weiteren Autor:innen. Die Arbeit wurde bisher 18-mal zitiert (Crossref) und nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Oasis Ecological Agriculture Corps Key Laboratory Open Project, Corps Science and Technology Program.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Hazy weather reduces contrast, narrows the dynamic range, and blurs the details of the remote sensing image. Additionally, color fidelity deteriorates, causing color shifts and image distortion, thereby impairing the utility of remote sensing data. In this paper, we propose a lightweight remote sensing-image-dehazing network, named LRSDN. The network comprises two tailored, lightweight modules arranged in cascade. The first module, the axial depthwise convolution and residual learning block (ADRB), is for feature extraction, efficiently expanding the convolutional receptive field with little computational overhead. The second is a feature-calibration module based on the hybrid attention block (HAB), which integrates a simplified, yet effective channel attention module and a pixel attention module embedded with an observational prior. This joint attention mechanism effectively enhances the representation of haze features. Furthermore, we introduce a novel method for remote sensing hazy image synthesis using Perlin noise, facilitating the creation of a large-scale, fine-grained remote sensing haze image dataset (RSHD). Finally, we conduct both quantitative and qualitative comparison experiments on multiple publicly available datasets. The results demonstrate that the LRSDN algorithm achieves superior dehazing performance with fewer than 0.1M parameters. We also validate the positive effects of the LRSDN in road extraction and land cover classification applications.

Förderung

  • National Natural Science Foundation of China
  • Oasis Ecological Agriculture Corps Key Laboratory Open Project
  • Corps Science and Technology Program
  • Innovation Team Project of Tarim University
  • Joint Funds of Tarim University and China Agricultural University

Zitieren

APA

He, Y., Li, C., Li, X., & Bai, T. (2024). A Lightweight CNN Based on Axial Depthwise Convolution and Hybrid Attention for Remote Sensing Image Dehazing. Remote Sensing, 16(15), 2822. https://doi.org/10.3390/rs16152822

Vancouver

He Y, Li C, Li X, Bai T. A Lightweight CNN Based on Axial Depthwise Convolution and Hybrid Attention for Remote Sensing Image Dehazing. Remote Sensing. 2024;16(15):2822. doi:10.3390/rs16152822

BibTeX
@article{He2024,
  author = {He, Yufeng and Li, Cuili and Li, Xu and Bai, Tiecheng},
  title = {A Lightweight CNN Based on Axial Depthwise Convolution and Hybrid Attention for Remote Sensing Image Dehazing},
  journal = {Remote Sensing},
  year = {2024},
  volume = {16},
  number = {15},
  pages = {2822},
  doi = {10.3390/rs16152822}
}
Kostenlos starten →