From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images
Accurate object detection is important in computer vision. However, detecting small objects in low-resolution images remains a challenging and elusive problem, primarily because these objects are constructed of less visual information and cannot be easily distinguished from similar background region...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-07-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/13/2620 |
| _version_ | 1827688149335146496 |
|---|---|
| author | Jingqian Wu Shibiao Xu |
| author_facet | Jingqian Wu Shibiao Xu |
| author_sort | Jingqian Wu |
| collection | DOAJ |
| description | Accurate object detection is important in computer vision. However, detecting small objects in low-resolution images remains a challenging and elusive problem, primarily because these objects are constructed of less visual information and cannot be easily distinguished from similar background regions. To resolve this problem, we propose a Hierarchical Small Object Detection Network in low-resolution remote sensing images, named HSOD-Net. We develop a point-to-region detection paradigm by first performing a key-point prediction to obtain position hypotheses, then only later super-resolving the image and detecting the objects around those candidate positions. By postponing the object prediction to after increasing its resolution, the obtained key-points are more stable than their traditional counterparts based on early object detection with less visual information. This hierarchical approach, HSOD-Net, saves significant run-time, which makes it more suitable for practical applications such as search and rescue, and drone navigation. In comparison with the state-of-art models, HSOD-Net achieves remarkable precision in detecting small objects in low-resolution remote sensing images. |
| first_indexed | 2024-03-10T09:49:52Z |
| format | Article |
| id | doaj.art-723b205c16d84c5e91145bd5267d15a1 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T09:49:52Z |
| publishDate | 2021-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-723b205c16d84c5e91145bd5267d15a12023-11-22T02:49:58ZengMDPI AGRemote Sensing2072-42922021-07-011313262010.3390/rs13132620From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing ImagesJingqian Wu0Shibiao Xu1Department of Computer Science, Wake Forest University, Winston-Salem, NC 27019, USAInstitute of Automation, Chinese Academy of Sciences, Beijing 100190, ChinaAccurate object detection is important in computer vision. However, detecting small objects in low-resolution images remains a challenging and elusive problem, primarily because these objects are constructed of less visual information and cannot be easily distinguished from similar background regions. To resolve this problem, we propose a Hierarchical Small Object Detection Network in low-resolution remote sensing images, named HSOD-Net. We develop a point-to-region detection paradigm by first performing a key-point prediction to obtain position hypotheses, then only later super-resolving the image and detecting the objects around those candidate positions. By postponing the object prediction to after increasing its resolution, the obtained key-points are more stable than their traditional counterparts based on early object detection with less visual information. This hierarchical approach, HSOD-Net, saves significant run-time, which makes it more suitable for practical applications such as search and rescue, and drone navigation. In comparison with the state-of-art models, HSOD-Net achieves remarkable precision in detecting small objects in low-resolution remote sensing images.https://www.mdpi.com/2072-4292/13/13/2620small object detectionkey-point predictionimage enhancementlow resolution |
| spellingShingle | Jingqian Wu Shibiao Xu From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images Remote Sensing small object detection key-point prediction image enhancement low resolution |
| title | From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images |
| title_full | From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images |
| title_fullStr | From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images |
| title_full_unstemmed | From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images |
| title_short | From Point to Region: Accurate and Efficient Hierarchical Small Object Detection in Low-Resolution Remote Sensing Images |
| title_sort | from point to region accurate and efficient hierarchical small object detection in low resolution remote sensing images |
| topic | small object detection key-point prediction image enhancement low resolution |
| url | https://www.mdpi.com/2072-4292/13/13/2620 |
| work_keys_str_mv | AT jingqianwu frompointtoregionaccurateandefficienthierarchicalsmallobjectdetectioninlowresolutionremotesensingimages AT shibiaoxu frompointtoregionaccurateandefficienthierarchicalsmallobjectdetectioninlowresolutionremotesensingimages |