MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data

Pinpointing elusive and minor target vessels from satellite-based images is recognized as a considerable obstacle in the specialized areas of computer vision and the examination of remote sensing imagery. The majority of existing methods are based on the YOLO architecture, which relies on manually d...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhuhua Hu, Xiyu Fan, Yaochi Zhao, Wei Wu, Jie Liu
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/11096611/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849228549970460672
author Zhuhua Hu
Xiyu Fan
Yaochi Zhao
Wei Wu
Jie Liu
author_facet Zhuhua Hu
Xiyu Fan
Yaochi Zhao
Wei Wu
Jie Liu
author_sort Zhuhua Hu
collection DOAJ
description Pinpointing elusive and minor target vessels from satellite-based images is recognized as a considerable obstacle in the specialized areas of computer vision and the examination of remote sensing imagery. The majority of existing methods are based on the YOLO architecture, which relies on manually designed anchor points and nonmaximum suppression (NMS) postprocessing. The detection of small targets in a single scene, the phenomenon of “catastrophic forgetting” due to the streaming of data, and the issue of an “information bottleneck” present significant challenges in this field. In order to address these issues, we propose the following solutions. A hybrid explicit spatial prior MH-Net network based on Manhattan distance is designed. By decomposing the self-attention matrix and the spatial attenuation matrix, the spatial correlations of different directions and positions are captured, thus effectively alleviating the problem of catastrophic forgetting. We propose an online convolutional reparameterization efficient layer aggregation networks cross-stage fusion network. Through equivalent transformations, the complex network architecture is compressed into a single linear layer. The network groups and processes input features in parallel, integrating both low-dimensional and high-dimensional features to alleviate the information bottleneck problem. The prediction head of the model uses the DINO decoder and applies contrastive denoising to remove useless prediction boxes. This allows the proposed MHOE-DETR model to avoid thresholding and NMS), reducing the model’s computational complexity. The experimental results demonstrate that the MHOE-DETR algorithm, designed for this purpose, markedly enhances the detection performance of small and indistinct targets in private remote sensing datasets. The average accuracy, recall, and AP50 reached 96.3%, 91.4%, and 95.4%, respectively, while maintaining a low GFLOPS value (54.4 G) and parametric count (77.3 M). These findings offer substantial technical justification for the implementation of sea area management and maritime safety monitoring strategies.
format Article
id doaj-art-e650580e9c564e71a4722b98b840b11e
institution Kabale University
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-e650580e9c564e71a4722b98b840b11e2025-08-22T23:08:06ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118204522046810.1109/JSTARS.2025.359285011096611MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image DataZhuhua Hu0https://orcid.org/0000-0002-6837-9024Xiyu Fan1Yaochi Zhao2https://orcid.org/0000-0003-0306-3114Wei Wu3https://orcid.org/0009-0002-5931-308XJie Liu4https://orcid.org/0009-0002-4550-9364School of Information and Communication Engineering, Hainan University, Haikou, ChinaSchool of Information and Communication Engineering, Hainan University, Haikou, ChinaSchool of Cyberspace Security, Hainan University, Haikou, ChinaSchool of Information and Communication Engineering, Hainan University, Haikou, ChinaSchool of Information and Communication Engineering, Hainan University, Haikou, ChinaPinpointing elusive and minor target vessels from satellite-based images is recognized as a considerable obstacle in the specialized areas of computer vision and the examination of remote sensing imagery. The majority of existing methods are based on the YOLO architecture, which relies on manually designed anchor points and nonmaximum suppression (NMS) postprocessing. The detection of small targets in a single scene, the phenomenon of “catastrophic forgetting” due to the streaming of data, and the issue of an “information bottleneck” present significant challenges in this field. In order to address these issues, we propose the following solutions. A hybrid explicit spatial prior MH-Net network based on Manhattan distance is designed. By decomposing the self-attention matrix and the spatial attenuation matrix, the spatial correlations of different directions and positions are captured, thus effectively alleviating the problem of catastrophic forgetting. We propose an online convolutional reparameterization efficient layer aggregation networks cross-stage fusion network. Through equivalent transformations, the complex network architecture is compressed into a single linear layer. The network groups and processes input features in parallel, integrating both low-dimensional and high-dimensional features to alleviate the information bottleneck problem. The prediction head of the model uses the DINO decoder and applies contrastive denoising to remove useless prediction boxes. This allows the proposed MHOE-DETR model to avoid thresholding and NMS), reducing the model’s computational complexity. The experimental results demonstrate that the MHOE-DETR algorithm, designed for this purpose, markedly enhances the detection performance of small and indistinct targets in private remote sensing datasets. The average accuracy, recall, and AP50 reached 96.3%, 91.4%, and 95.4%, respectively, while maintaining a low GFLOPS value (54.4 G) and parametric count (77.3 M). These findings offer substantial technical justification for the implementation of sea area management and maritime safety monitoring strategies.https://ieeexplore.ieee.org/document/11096611/Fuzzy targetsinformation bottlenecksmodel catastrophic forgetremote sensing datasmall targets
spellingShingle Zhuhua Hu
Xiyu Fan
Yaochi Zhao
Wei Wu
Jie Liu
MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Fuzzy targets
information bottlenecks
model catastrophic forget
remote sensing data
small targets
title MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data
title_full MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data
title_fullStr MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data
title_full_unstemmed MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data
title_short MHOE-DETR: A Ship Detection Method for Small and Fuzzy Targets Based on Satellite Remote Sensing Image Data
title_sort mhoe detr a ship detection method for small and fuzzy targets based on satellite remote sensing image data
topic Fuzzy targets
information bottlenecks
model catastrophic forget
remote sensing data
small targets
url https://ieeexplore.ieee.org/document/11096611/
work_keys_str_mv AT zhuhuahu mhoedetrashipdetectionmethodforsmallandfuzzytargetsbasedonsatelliteremotesensingimagedata
AT xiyufan mhoedetrashipdetectionmethodforsmallandfuzzytargetsbasedonsatelliteremotesensingimagedata
AT yaochizhao mhoedetrashipdetectionmethodforsmallandfuzzytargetsbasedonsatelliteremotesensingimagedata
AT weiwu mhoedetrashipdetectionmethodforsmallandfuzzytargetsbasedonsatelliteremotesensingimagedata
AT jieliu mhoedetrashipdetectionmethodforsmallandfuzzytargetsbasedonsatelliteremotesensingimagedata