SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching

Due to the distant acquisition and bird’s-eye perspective of remote sensing images, ground objects are distributed in arbitrary scales and multiple orientations. Existing detectors often utilize feature pyramid networks (FPN) and deformable (or rotated) convolutions to adapt to variations in object...

Full description

Saved in:
Bibliographic Details
Main Authors: Keliang Liu, Yantao Xi, Donglin Jing, Xue Zhang, Mingfei Xu
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/9/1622
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850030304634339328
author Keliang Liu
Yantao Xi
Donglin Jing
Xue Zhang
Mingfei Xu
author_facet Keliang Liu
Yantao Xi
Donglin Jing
Xue Zhang
Mingfei Xu
author_sort Keliang Liu
collection DOAJ
description Due to the distant acquisition and bird’s-eye perspective of remote sensing images, ground objects are distributed in arbitrary scales and multiple orientations. Existing detectors often utilize feature pyramid networks (FPN) and deformable (or rotated) convolutions to adapt to variations in object scale and orientation. However, these methods solve scale and orientation issues separately and ignore their deeper coupling relationships. When the scale features extracted by the network are significantly mismatched with the object, it is difficult for the detection head to effectively capture orientation of object, resulting in misalignment between object and bounding box. Therefore, we propose a one-stage detector—Scale First Refinement-Angle Detection Network (SFRADNet), which aims to fine-tune the rotation angle under precise scale feature matching. We introduce the Group Learning Large Kernel Network (GL<sup>2</sup>KNet) as the backbone of SFRADNet and employ a Shape-Aware Spatial Feature Extraction Module (SA-SFEM) as the primary component of the detection head. Specifically, within GL<sup>2</sup>KNet, we construct diverse receptive fields with varying dilation rates to capture features across different spatial coverage ranges. Building on this, we utilize multi-scale features within the layers and apply weighted aggregation based on a Scale Selection Matrix (SSMatrix). The SSMatrix dynamically adjusts the receptive field coverage according to the target size, enabling more refined selection of scale features. Based on precise scale features captured, we first design a Directed Guiding Box (DGBox) within the SA-SFEM, using its shape and position information to supervise the sampling points of the convolution kernels, thereby fitting them to deformations of object. This facilitates the extraction of orientation features near the object region, allowing for accurate refinement of both scale and orientation. Experiments show that our network achieves a mAP of 80.10% on the DOTA-v1.0 dataset, while reducing computational complexity compared to the baseline model.
format Article
id doaj-art-cefec97091fc4d3b903b6889b93bcf81
institution DOAJ
issn 2072-4292
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-cefec97091fc4d3b903b6889b93bcf812025-08-20T02:59:15ZengMDPI AGRemote Sensing2072-42922025-05-01179162210.3390/rs17091622SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature MatchingKeliang Liu0Yantao Xi1Donglin Jing2Xue Zhang3Mingfei Xu4School of Resources and GEOSciences, China University of Mining and Technology, Xuzhou 221116, ChinaSchool of Resources and GEOSciences, China University of Mining and Technology, Xuzhou 221116, ChinaSchool of Information and Electronics, Beijing Institute of Technology, Beijing 100081, ChinaSchool of Resources and GEOSciences, China University of Mining and Technology, Xuzhou 221116, ChinaSchool of Resources and GEOSciences, China University of Mining and Technology, Xuzhou 221116, ChinaDue to the distant acquisition and bird’s-eye perspective of remote sensing images, ground objects are distributed in arbitrary scales and multiple orientations. Existing detectors often utilize feature pyramid networks (FPN) and deformable (or rotated) convolutions to adapt to variations in object scale and orientation. However, these methods solve scale and orientation issues separately and ignore their deeper coupling relationships. When the scale features extracted by the network are significantly mismatched with the object, it is difficult for the detection head to effectively capture orientation of object, resulting in misalignment between object and bounding box. Therefore, we propose a one-stage detector—Scale First Refinement-Angle Detection Network (SFRADNet), which aims to fine-tune the rotation angle under precise scale feature matching. We introduce the Group Learning Large Kernel Network (GL<sup>2</sup>KNet) as the backbone of SFRADNet and employ a Shape-Aware Spatial Feature Extraction Module (SA-SFEM) as the primary component of the detection head. Specifically, within GL<sup>2</sup>KNet, we construct diverse receptive fields with varying dilation rates to capture features across different spatial coverage ranges. Building on this, we utilize multi-scale features within the layers and apply weighted aggregation based on a Scale Selection Matrix (SSMatrix). The SSMatrix dynamically adjusts the receptive field coverage according to the target size, enabling more refined selection of scale features. Based on precise scale features captured, we first design a Directed Guiding Box (DGBox) within the SA-SFEM, using its shape and position information to supervise the sampling points of the convolution kernels, thereby fitting them to deformations of object. This facilitates the extraction of orientation features near the object region, allowing for accurate refinement of both scale and orientation. Experiments show that our network achieves a mAP of 80.10% on the DOTA-v1.0 dataset, while reducing computational complexity compared to the baseline model.https://www.mdpi.com/2072-4292/17/9/1622remote sensing imageone-stage detectordeformable convolutionsscale featureorientation feature
spellingShingle Keliang Liu
Yantao Xi
Donglin Jing
Xue Zhang
Mingfei Xu
SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching
Remote Sensing
remote sensing image
one-stage detector
deformable convolutions
scale feature
orientation feature
title SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching
title_full SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching
title_fullStr SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching
title_full_unstemmed SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching
title_short SFRADNet: Object Detection Network with Angle Fine-Tuning Under Feature Matching
title_sort sfradnet object detection network with angle fine tuning under feature matching
topic remote sensing image
one-stage detector
deformable convolutions
scale feature
orientation feature
url https://www.mdpi.com/2072-4292/17/9/1622
work_keys_str_mv AT keliangliu sfradnetobjectdetectionnetworkwithanglefinetuningunderfeaturematching
AT yantaoxi sfradnetobjectdetectionnetworkwithanglefinetuningunderfeaturematching
AT donglinjing sfradnetobjectdetectionnetworkwithanglefinetuningunderfeaturematching
AT xuezhang sfradnetobjectdetectionnetworkwithanglefinetuningunderfeaturematching
AT mingfeixu sfradnetobjectdetectionnetworkwithanglefinetuningunderfeaturematching