KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection

Spaceborne infrared imagery provides substantial data support for ship detection due to its adaptability and all-weather reconnaissance capabilities. However, in complex maritime environments, its inadequate semantic information and feature distinctiveness increase the risk of false alarms and misse...

Full description

Saved in:
Bibliographic Details
Main Authors: Yifei Peng, Ze Tao, Jian Zhang, Rongchang Zhao, Chao Liu, Hui Sun
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11072708/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850067648048529408
author Yifei Peng
Ze Tao
Jian Zhang
Rongchang Zhao
Chao Liu
Hui Sun
author_facet Yifei Peng
Ze Tao
Jian Zhang
Rongchang Zhao
Chao Liu
Hui Sun
author_sort Yifei Peng
collection DOAJ
description Spaceborne infrared imagery provides substantial data support for ship detection due to its adaptability and all-weather reconnaissance capabilities. However, in complex maritime environments, its inadequate semantic information and feature distinctiveness increase the risk of false alarms and missed detections. To address these challenges, we propose a knowledge-driven feature fusion detection transformer (KDFF-DETR) for spaceborne infrared ship detection. KDFF-DETR presents an enhancement of RT-DETR. Specifically, we design a hybrid Fourier contextual (HFC) encoder to model the adequate semantic representation of high-level features through frequency-domain features and masked supervision. We further develop a visual knowledge-supervised feature reconstruction (VKFR) module, which introduces feature operators to represent visual knowledge and reconstruct handcrafted features to enhance their distinctiveness. Besides, a contribution-balanced feature fusion (CBFF) method for query selection is proposed, dynamically adjusting the weight distribution of handcrafted and deep features during different training stages to improve feature fusion accuracy and generalization ability. Experimental results indicate that KDFF-DETR surpasses state-of-the-art (SOTA) methods in various public datasets.
format Article
id doaj-art-6491b2fd57a141bbbd296767d8f721f6
institution DOAJ
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-6491b2fd57a141bbbd296767d8f721f62025-08-20T02:48:16ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118176381765310.1109/JSTARS.2025.358695511072708KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship DetectionYifei Peng0https://orcid.org/0009-0009-8951-0106Ze Tao1https://orcid.org/0000-0003-4009-194XJian Zhang2https://orcid.org/0000-0001-5418-0455Rongchang Zhao3https://orcid.org/0000-0002-5171-4121Chao Liu4Hui Sun5https://orcid.org/0009-0001-3713-1982School of Computer Science and Engineering, Central South University, Changsha, ChinaSchool of Computer Science and Engineering, Central South University, Changsha, ChinaSchool of Computer Science and Engineering, Central South University, Changsha, ChinaSchool of Computer Science and Engineering, Central South University, Changsha, ChinaInstitute of Systems Engineering, Academy of Military Science, Beijing, ChinaInstitute of Systems Engineering, Academy of Military Science, Beijing, ChinaSpaceborne infrared imagery provides substantial data support for ship detection due to its adaptability and all-weather reconnaissance capabilities. However, in complex maritime environments, its inadequate semantic information and feature distinctiveness increase the risk of false alarms and missed detections. To address these challenges, we propose a knowledge-driven feature fusion detection transformer (KDFF-DETR) for spaceborne infrared ship detection. KDFF-DETR presents an enhancement of RT-DETR. Specifically, we design a hybrid Fourier contextual (HFC) encoder to model the adequate semantic representation of high-level features through frequency-domain features and masked supervision. We further develop a visual knowledge-supervised feature reconstruction (VKFR) module, which introduces feature operators to represent visual knowledge and reconstruct handcrafted features to enhance their distinctiveness. Besides, a contribution-balanced feature fusion (CBFF) method for query selection is proposed, dynamically adjusting the weight distribution of handcrafted and deep features during different training stages to improve feature fusion accuracy and generalization ability. Experimental results indicate that KDFF-DETR surpasses state-of-the-art (SOTA) methods in various public datasets.https://ieeexplore.ieee.org/document/11072708/Contribution-balanced feature fusion (CBFF)handcrafted feature reconstructionhigh-level semantic modelingknowledge-driven detection transformer
spellingShingle Yifei Peng
Ze Tao
Jian Zhang
Rongchang Zhao
Chao Liu
Hui Sun
KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Contribution-balanced feature fusion (CBFF)
handcrafted feature reconstruction
high-level semantic modeling
knowledge-driven detection transformer
title KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection
title_full KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection
title_fullStr KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection
title_full_unstemmed KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection
title_short KDFF-DETR: Knowledge-Driven Feature Fusion DETR for Spaceborne Infrared Ship Detection
title_sort kdff detr knowledge driven feature fusion detr for spaceborne infrared ship detection
topic Contribution-balanced feature fusion (CBFF)
handcrafted feature reconstruction
high-level semantic modeling
knowledge-driven detection transformer
url https://ieeexplore.ieee.org/document/11072708/
work_keys_str_mv AT yifeipeng kdffdetrknowledgedrivenfeaturefusiondetrforspaceborneinfraredshipdetection
AT zetao kdffdetrknowledgedrivenfeaturefusiondetrforspaceborneinfraredshipdetection
AT jianzhang kdffdetrknowledgedrivenfeaturefusiondetrforspaceborneinfraredshipdetection
AT rongchangzhao kdffdetrknowledgedrivenfeaturefusiondetrforspaceborneinfraredshipdetection
AT chaoliu kdffdetrknowledgedrivenfeaturefusiondetrforspaceborneinfraredshipdetection
AT huisun kdffdetrknowledgedrivenfeaturefusiondetrforspaceborneinfraredshipdetection