Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context
With the rapid development of deep learning technology in recent years, convolutional neural networks have gained remarkable progress in SAR ship detection tasks. However, noise interference of the background and inadequate appearance features of small-scale objects still pose challenges. To tackle...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Smart Cities |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2624-6511/6/3/76 |
| _version_ | 1797592680792850432 |
|---|---|
| author | Zhengjie Jiang Yupei Wang Xiaoqi Zhou Liang Chen Yuan Chang Dongsheng Song Hao Shi |
| author_facet | Zhengjie Jiang Yupei Wang Xiaoqi Zhou Liang Chen Yuan Chang Dongsheng Song Hao Shi |
| author_sort | Zhengjie Jiang |
| collection | DOAJ |
| description | With the rapid development of deep learning technology in recent years, convolutional neural networks have gained remarkable progress in SAR ship detection tasks. However, noise interference of the background and inadequate appearance features of small-scale objects still pose challenges. To tackle these issues, we propose a small ship detection algorithm for SAR images by means of a coordinate-aware mixed attention mechanism and spatial semantic joint context method. First, the coordinate-aware mixed attention mechanism innovatively combines coordinate-aware channel attention and spatial attention to achieve coordinate alignment of mixed attention features. In this way, attention with finer spatial granularity is conducive to strengthening the focusing ability on small-scale objects, thereby suppressing the background clutters accurately. In addition, the spatial semantic joint context method exploits the local and global environmental information jointly. The detailed spatial cues contained in the multi-scale local context and the generalized semantic information encoded in the global context are used to enhance the feature expression and distinctiveness of small-scale ship objects. Extensive experiments are conducted on the LS-SSDD-v1.0 and the HRSID dataset. The results with an average precision of 77.23% and 90.85% on the two datasets show the effectiveness of the proposed methods. |
| first_indexed | 2024-03-11T01:56:34Z |
| format | Article |
| id | doaj.art-f9091eac21cc4d38a7d6379712eecd06 |
| institution | Directory Open Access Journal |
| issn | 2624-6511 |
| language | English |
| last_indexed | 2024-03-11T01:56:34Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Smart Cities |
| spelling | doaj.art-f9091eac21cc4d38a7d6379712eecd062023-11-18T12:37:04ZengMDPI AGSmart Cities2624-65112023-06-01631612162910.3390/smartcities6030076Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint ContextZhengjie Jiang0Yupei Wang1Xiaoqi Zhou2Liang Chen3Yuan Chang4Dongsheng Song5Hao Shi6School of Information and Electronics, Beijing Institute of Technology (BIT), Beijing 100081, ChinaSchool of Information and Electronics, Beijing Institute of Technology (BIT), Beijing 100081, ChinaChongqing Innovation Center, Beijing Institute of Technology, Chongqing 401135, ChinaSchool of Information and Electronics, Beijing Institute of Technology (BIT), Beijing 100081, ChinaSchool of Information and Electronics, Beijing Institute of Technology (BIT), Beijing 100081, ChinaSchool of Information and Electronics, Beijing Institute of Technology (BIT), Beijing 100081, ChinaSchool of Information and Electronics, Beijing Institute of Technology (BIT), Beijing 100081, ChinaWith the rapid development of deep learning technology in recent years, convolutional neural networks have gained remarkable progress in SAR ship detection tasks. However, noise interference of the background and inadequate appearance features of small-scale objects still pose challenges. To tackle these issues, we propose a small ship detection algorithm for SAR images by means of a coordinate-aware mixed attention mechanism and spatial semantic joint context method. First, the coordinate-aware mixed attention mechanism innovatively combines coordinate-aware channel attention and spatial attention to achieve coordinate alignment of mixed attention features. In this way, attention with finer spatial granularity is conducive to strengthening the focusing ability on small-scale objects, thereby suppressing the background clutters accurately. In addition, the spatial semantic joint context method exploits the local and global environmental information jointly. The detailed spatial cues contained in the multi-scale local context and the generalized semantic information encoded in the global context are used to enhance the feature expression and distinctiveness of small-scale ship objects. Extensive experiments are conducted on the LS-SSDD-v1.0 and the HRSID dataset. The results with an average precision of 77.23% and 90.85% on the two datasets show the effectiveness of the proposed methods.https://www.mdpi.com/2624-6511/6/3/76Synthetic Aperture Radar (SAR)ship detectionsmall object detectionattention mechanismcontext |
| spellingShingle | Zhengjie Jiang Yupei Wang Xiaoqi Zhou Liang Chen Yuan Chang Dongsheng Song Hao Shi Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context Smart Cities Synthetic Aperture Radar (SAR) ship detection small object detection attention mechanism context |
| title | Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context |
| title_full | Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context |
| title_fullStr | Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context |
| title_full_unstemmed | Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context |
| title_short | Small-Scale Ship Detection for SAR Remote Sensing Images Based on Coordinate-Aware Mixed Attention and Spatial Semantic Joint Context |
| title_sort | small scale ship detection for sar remote sensing images based on coordinate aware mixed attention and spatial semantic joint context |
| topic | Synthetic Aperture Radar (SAR) ship detection small object detection attention mechanism context |
| url | https://www.mdpi.com/2624-6511/6/3/76 |
| work_keys_str_mv | AT zhengjiejiang smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext AT yupeiwang smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext AT xiaoqizhou smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext AT liangchen smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext AT yuanchang smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext AT dongshengsong smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext AT haoshi smallscaleshipdetectionforsarremotesensingimagesbasedoncoordinateawaremixedattentionandspatialsemanticjointcontext |