RFWNet: A Lightweight Remote Sensing Object Detector Integrating Multi-Scale Receptive Fields and Foreground Focus Mechanism

Challenges in remote sensing object detection (RSOD), such as high inter-class similarity, imbalanced foreground-background distribution, and the small size of objects in remote sensing images significantly hinder detection accuracy. Moreo-ver, the trade-off between model accuracy and computational complexity poses additional constraints on the application of RSOD algorithms. To address these issues, this study proposes an efficient and lightweight RSOD algorithm integrat-ing multi-scale receptive fields and foreground focus mechanism, named RFWNet. Specifically, we proposed a lightweight backbone network Receptive Field Adaptive Selection Network (RFASNet), leveraging the rich context infor-mation of remote sensing images to enhance class separability. Additionally, we developed a Foreground Background Separation Module (FBSM) consisting of a background redundant information filtering module and a foreground information enhancement module to emphasize critical regions within images while filtering redundant background information. Finally, we designed a loss function, the Weighted CIoU-Wasserstein (WCW) loss, which weights the IoU-based loss by using the Normalized Wasserstein Distance to mitigate model sensitivity to small object position deviations. Experimental evaluations on the DOTA V1.0 and NWPU VHR-10 datasets demonstrate that RFWNet achieves advanced perfor-mance with 6.0M parameters and can achieves 52 FPS.

@article{lei2025_2503.00545,
  title={ RFWNet: A Lightweight Remote Sensing Object Detector Integrating Multi-Scale Receptive Fields and Foreground Focus Mechanism },
  author={ Yujie Lei and Wenjie Sun and Sen Jia and Qingquan Li and Jie Zhang },
  journal={arXiv preprint arXiv:2503.00545},
  year={ 2025 }
}