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

2025-03-01

AI Summary

Remote sensing imagery small-object detection faces challenges including high inter-class similarity, extreme foreground-background imbalance, and difficulty in model lightweighting. To address these, we propose RFASNet—a lightweight backbone network featuring multi-scale receptive field adaptive selection—along with a Foreground-Focusing Separation Module (FBSM) that filters background redundancy while enhancing foreground features. Additionally, we introduce a composite loss function, Weighted CIoU-Wasserstein (WCW), to jointly mitigate distribution shift and resolve conflicts between classification and localization optimization. Evaluated on DOTA v1.0 and NWPU VHR-10, our method achieves state-of-the-art accuracy with only 6.0M parameters and 52 FPS inference speed, significantly improving both robustness and real-time performance for small-object detection in remote sensing imagery.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

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.

Problem

• Addresses high inter-class similarity in remote sensing object detection.
• Improves detection of small objects in remote sensing images.
• Balances model accuracy and computational complexity for efficiency.

Innovation

• Lightweight backbone network RFASNet enhances class separability.
• Foreground Background Separation Module emphasizes critical image regions.
• Weighted CIoU-Wasserstein loss mitigates small object position sensitivity.