DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts
Aircraft target detection in remote sensing images faces numerous challenges, including target size variations, low resolution, and complex backgrounds. To address these challenges, an enhanced end-to-end aircraft detection framework (DIMD-DETR) is developed based on an improved metric space. Initia...
Saved in:
| Main Authors: | , , , , |
|---|---|
| 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/10843752/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832542559485820928 |
|---|---|
| author | Huan Liu Xuefeng Ren Yang Gan Yongming Chen Ping Lin |
| author_facet | Huan Liu Xuefeng Ren Yang Gan Yongming Chen Ping Lin |
| author_sort | Huan Liu |
| collection | DOAJ |
| description | Aircraft target detection in remote sensing images faces numerous challenges, including target size variations, low resolution, and complex backgrounds. To address these challenges, an enhanced end-to-end aircraft detection framework (DIMD-DETR) is developed based on an improved metric space. Initially, a bilayer targeted prediction method is proposed to strengthen gradient interaction across decoder layers, thereby enhancing detection accuracy and sensitivity in complex scenarios. The pyramid structure and self-attention mechanism from pyramid vision transformer V2 are incorporated to enable effective joint learning of both global and local features, which significantly boosts performance for low-resolution targets. To further enhance the model's generalization capabilities, an aircraft-specific data augmentation strategy is meticulously devised, thereby improving the model's adaptability to variations in scale and appearance. In addition, a metric-space-based loss function is developed to optimize the collaborative effects of the modular architecture, enhancing detection performance in complex backgrounds and under varying target conditions. Finally, a dynamic learning rate scheduling strategy is proposed to balance rapid convergence with global exploration, thereby elevating the model's robustness in challenging environments. Compared to current popular networks, our model demonstrated superior detection performance with fewer parameters. |
| format | Article |
| id | doaj-art-097546c6ec5047cabe4d8145fa8fc293 |
| 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-097546c6ec5047cabe4d8145fa8fc2932025-02-04T00:00:26ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01184498450910.1109/JSTARS.2025.353014110843752DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing AircraftsHuan Liu0Xuefeng Ren1Yang Gan2Yongming Chen3Ping Lin4https://orcid.org/0009-0008-9946-8434School of Electrical Engineering and Automation, Hubei Normal University, Huangshi, ChinaSchool of Electrical Engineering and Automation, Hubei Normal University, Huangshi, ChinaSchool of Electrical Engineering and Automation, Hubei Normal University, Huangshi, ChinaSchool of Electrical Engineering and Automation, Hubei Normal University, Huangshi, ChinaSchool of Electrical Engineering and Automation, Hubei Normal University, Huangshi, ChinaAircraft target detection in remote sensing images faces numerous challenges, including target size variations, low resolution, and complex backgrounds. To address these challenges, an enhanced end-to-end aircraft detection framework (DIMD-DETR) is developed based on an improved metric space. Initially, a bilayer targeted prediction method is proposed to strengthen gradient interaction across decoder layers, thereby enhancing detection accuracy and sensitivity in complex scenarios. The pyramid structure and self-attention mechanism from pyramid vision transformer V2 are incorporated to enable effective joint learning of both global and local features, which significantly boosts performance for low-resolution targets. To further enhance the model's generalization capabilities, an aircraft-specific data augmentation strategy is meticulously devised, thereby improving the model's adaptability to variations in scale and appearance. In addition, a metric-space-based loss function is developed to optimize the collaborative effects of the modular architecture, enhancing detection performance in complex backgrounds and under varying target conditions. Finally, a dynamic learning rate scheduling strategy is proposed to balance rapid convergence with global exploration, thereby elevating the model's robustness in challenging environments. Compared to current popular networks, our model demonstrated superior detection performance with fewer parameters.https://ieeexplore.ieee.org/document/10843752/Aircraft detectionend-to-endmetric spaceremote sensingtransformer |
| spellingShingle | Huan Liu Xuefeng Ren Yang Gan Yongming Chen Ping Lin DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Aircraft detection end-to-end metric space remote sensing transformer |
| title | DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts |
| title_full | DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts |
| title_fullStr | DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts |
| title_full_unstemmed | DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts |
| title_short | DIMD-DETR: DDQ-DETR With Improved Metric Space for End-to-End Object Detector on Remote Sensing Aircrafts |
| title_sort | dimd detr ddq detr with improved metric space for end to end object detector on remote sensing aircrafts |
| topic | Aircraft detection end-to-end metric space remote sensing transformer |
| url | https://ieeexplore.ieee.org/document/10843752/ |
| work_keys_str_mv | AT huanliu dimddetrddqdetrwithimprovedmetricspaceforendtoendobjectdetectoronremotesensingaircrafts AT xuefengren dimddetrddqdetrwithimprovedmetricspaceforendtoendobjectdetectoronremotesensingaircrafts AT yanggan dimddetrddqdetrwithimprovedmetricspaceforendtoendobjectdetectoronremotesensingaircrafts AT yongmingchen dimddetrddqdetrwithimprovedmetricspaceforendtoendobjectdetectoronremotesensingaircrafts AT pinglin dimddetrddqdetrwithimprovedmetricspaceforendtoendobjectdetectoronremotesensingaircrafts |