Priority Branches for Ship Detection in Optical Remote Sensing Images
Much attention is being paid to using high-performance convolutional neural networks (CNNs) in the area of ship detection in optical remoting sensing (ORS) images. However, the problem of false negatives (FNs) caused by side-by-side ships cannot be solved, and the number of false positives (FPs) rem...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/7/1196 |
| _version_ | 1797571221813985280 |
|---|---|
| author | Yijia Zhang Weiguang Sheng Jianfei Jiang Naifeng Jing Qin Wang Zhigang Mao |
| author_facet | Yijia Zhang Weiguang Sheng Jianfei Jiang Naifeng Jing Qin Wang Zhigang Mao |
| author_sort | Yijia Zhang |
| collection | DOAJ |
| description | Much attention is being paid to using high-performance convolutional neural networks (CNNs) in the area of ship detection in optical remoting sensing (ORS) images. However, the problem of false negatives (FNs) caused by side-by-side ships cannot be solved, and the number of false positives (FPs) remains high. This paper uses a DLA-34 network with deformable convolution layers as the backbone. The network has two priority branches: a recall-priority branch for reducing the number of FNs, and a precision-priority branch for reducing the number of FPs. In our single-shot detection method, the recall-priority branch is based on an anchor-free module without non-maximum suppression (NMS), while the precision-priority branch utilizes an anchor-based module with NMS. We perform recall-priority branch functions based on the output part of the CenterNet object detector to precisely predict center points of bounding boxes. The Bidirectional Feature Pyramid Network (BiFPN), combined with the inference part of YOLOv3, is used to improve the precision of precision-priority branch. Finally, the boxes from two branches merge, and we propose priority-based selection (PBS) for choosing the accurate ones. Results show that our proposed method sharply improves the recall rate of side-by-side ships and significantly reduces the number of false alarms. Our method also achieves the best trade-off on our improved version of HRSC2016 dataset, with 95.57% AP at 56 frames per second on an Nvidia RTX-2080 Ti GPU. Compared with the HRSC2016 dataset, not only are our annotations more accurate, but our dataset also contains more images and samples. Our evaluation metrics also included tests on small ships and incomplete forms of ships. |
| first_indexed | 2024-03-10T20:36:17Z |
| format | Article |
| id | doaj.art-dd0b649f99e54c66820cdc4a986e29bd |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T20:36:17Z |
| publishDate | 2020-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-dd0b649f99e54c66820cdc4a986e29bd2023-11-19T21:01:27ZengMDPI AGRemote Sensing2072-42922020-04-01127119610.3390/rs12071196Priority Branches for Ship Detection in Optical Remote Sensing ImagesYijia Zhang0Weiguang Sheng1Jianfei Jiang2Naifeng Jing3Qin Wang4Zhigang Mao5Department of Micro/Nano Electronics, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaDepartment of Micro/Nano Electronics, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaDepartment of Micro/Nano Electronics, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaDepartment of Micro/Nano Electronics, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaDepartment of Micro/Nano Electronics, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaDepartment of Micro/Nano Electronics, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, ChinaMuch attention is being paid to using high-performance convolutional neural networks (CNNs) in the area of ship detection in optical remoting sensing (ORS) images. However, the problem of false negatives (FNs) caused by side-by-side ships cannot be solved, and the number of false positives (FPs) remains high. This paper uses a DLA-34 network with deformable convolution layers as the backbone. The network has two priority branches: a recall-priority branch for reducing the number of FNs, and a precision-priority branch for reducing the number of FPs. In our single-shot detection method, the recall-priority branch is based on an anchor-free module without non-maximum suppression (NMS), while the precision-priority branch utilizes an anchor-based module with NMS. We perform recall-priority branch functions based on the output part of the CenterNet object detector to precisely predict center points of bounding boxes. The Bidirectional Feature Pyramid Network (BiFPN), combined with the inference part of YOLOv3, is used to improve the precision of precision-priority branch. Finally, the boxes from two branches merge, and we propose priority-based selection (PBS) for choosing the accurate ones. Results show that our proposed method sharply improves the recall rate of side-by-side ships and significantly reduces the number of false alarms. Our method also achieves the best trade-off on our improved version of HRSC2016 dataset, with 95.57% AP at 56 frames per second on an Nvidia RTX-2080 Ti GPU. Compared with the HRSC2016 dataset, not only are our annotations more accurate, but our dataset also contains more images and samples. Our evaluation metrics also included tests on small ships and incomplete forms of ships.https://www.mdpi.com/2072-4292/12/7/1196ship detectionoptical remote sensing imagespriority branchside-by-side ships |
| spellingShingle | Yijia Zhang Weiguang Sheng Jianfei Jiang Naifeng Jing Qin Wang Zhigang Mao Priority Branches for Ship Detection in Optical Remote Sensing Images Remote Sensing ship detection optical remote sensing images priority branch side-by-side ships |
| title | Priority Branches for Ship Detection in Optical Remote Sensing Images |
| title_full | Priority Branches for Ship Detection in Optical Remote Sensing Images |
| title_fullStr | Priority Branches for Ship Detection in Optical Remote Sensing Images |
| title_full_unstemmed | Priority Branches for Ship Detection in Optical Remote Sensing Images |
| title_short | Priority Branches for Ship Detection in Optical Remote Sensing Images |
| title_sort | priority branches for ship detection in optical remote sensing images |
| topic | ship detection optical remote sensing images priority branch side-by-side ships |
| url | https://www.mdpi.com/2072-4292/12/7/1196 |
| work_keys_str_mv | AT yijiazhang prioritybranchesforshipdetectioninopticalremotesensingimages AT weiguangsheng prioritybranchesforshipdetectioninopticalremotesensingimages AT jianfeijiang prioritybranchesforshipdetectioninopticalremotesensingimages AT naifengjing prioritybranchesforshipdetectioninopticalremotesensingimages AT qinwang prioritybranchesforshipdetectioninopticalremotesensingimages AT zhigangmao prioritybranchesforshipdetectioninopticalremotesensingimages |