Ship Detection Based on YOLOv2 for SAR Imagery
Synthetic aperture radar (SAR) imagery has been used as a promising data source for monitoring maritime activities, and its application for oil and ship detection has been the focus of many previous research studies. Many object detection methods ranging from traditional to deep learning approaches...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/11/7/786 |
| _version_ | 1798027054056210432 |
|---|---|
| author | Yang-Lang Chang Amare Anagaw Lena Chang Yi Chun Wang Chih-Yu Hsiao Wei-Hong Lee |
| author_facet | Yang-Lang Chang Amare Anagaw Lena Chang Yi Chun Wang Chih-Yu Hsiao Wei-Hong Lee |
| author_sort | Yang-Lang Chang |
| collection | DOAJ |
| description | Synthetic aperture radar (SAR) imagery has been used as a promising data source for monitoring maritime activities, and its application for oil and ship detection has been the focus of many previous research studies. Many object detection methods ranging from traditional to deep learning approaches have been proposed. However, majority of them are computationally intensive and have accuracy problems. The huge volume of the remote sensing data also brings a challenge for real time object detection. To mitigate this problem a <i>high performance computing</i> (HPC) method has been proposed to accelerate SAR imagery analysis, utilizing the GPU based computing methods. In this paper, we propose an enhanced GPU based deep learning method to detect ship from the SAR images. The <i>You Only Look Once version 2</i> (YOLOv2) deep learning framework is proposed to model the architecture and training the model. YOLOv2 is a state-of-the-art real-time object detection system, which outperforms <i>Faster Region-Based Convolutional Network</i> (Faster R-CNN) and <i>Single Shot Multibox Detector</i> (SSD) methods. Additionally, in order to reduce computational time with relatively competitive detection accuracy, we develop a new architecture with less number of layers called <i>YOLOv2-reduced</i>. In the experiment, we use two types of datasets: A <i>SAR ship detection dataset</i> (SSDD) dataset and a <i>Diversified SAR Ship Detection Dataset</i> (DSSDD). These two datasets were used for training and testing purposes. YOLOv2 test results showed an increase in accuracy of ship detection as well as a noticeable reduction in computational time compared to Faster R-CNN. From the experimental results, the proposed YOLOv2 architecture achieves an accuracy of 90.05% and 89.13% on the SSDD and DSSDD datasets respectively. The proposed <i>YOLOv2-reduced</i> architecture has a similarly competent detection performance as YOLOv2, but with less computational time on a NVIDIA TITAN X GPU. The experimental results shows that the deep learning can make a big leap forward in improving the performance of SAR image ship detection. |
| first_indexed | 2024-04-11T18:45:05Z |
| format | Article |
| id | doaj.art-5304c5b4d141423fb03ea61f4fb2fd4c |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-04-11T18:45:05Z |
| publishDate | 2019-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-5304c5b4d141423fb03ea61f4fb2fd4c2022-12-22T04:08:50ZengMDPI AGRemote Sensing2072-42922019-04-0111778610.3390/rs11070786rs11070786Ship Detection Based on YOLOv2 for SAR ImageryYang-Lang Chang0Amare Anagaw1Lena Chang2Yi Chun Wang3Chih-Yu Hsiao4Wei-Hong Lee5Department of Electrical Engineering, National Taipei University of Technology, Taipei 10608, TaiwanDepartment of Electrical Engineering, National Taipei University of Technology, Taipei 10608, TaiwanDepartment of Communications and Guidance Engineering, National Taiwan Ocean University, Keelung 20248, TaiwanDepartment of Electrical Engineering, National Taipei University of Technology, Taipei 10608, TaiwanDepartment of Electrical Engineering, National Taipei University of Technology, Taipei 10608, TaiwanDepartment of Electrical Engineering, National Taipei University of Technology, Taipei 10608, TaiwanSynthetic aperture radar (SAR) imagery has been used as a promising data source for monitoring maritime activities, and its application for oil and ship detection has been the focus of many previous research studies. Many object detection methods ranging from traditional to deep learning approaches have been proposed. However, majority of them are computationally intensive and have accuracy problems. The huge volume of the remote sensing data also brings a challenge for real time object detection. To mitigate this problem a <i>high performance computing</i> (HPC) method has been proposed to accelerate SAR imagery analysis, utilizing the GPU based computing methods. In this paper, we propose an enhanced GPU based deep learning method to detect ship from the SAR images. The <i>You Only Look Once version 2</i> (YOLOv2) deep learning framework is proposed to model the architecture and training the model. YOLOv2 is a state-of-the-art real-time object detection system, which outperforms <i>Faster Region-Based Convolutional Network</i> (Faster R-CNN) and <i>Single Shot Multibox Detector</i> (SSD) methods. Additionally, in order to reduce computational time with relatively competitive detection accuracy, we develop a new architecture with less number of layers called <i>YOLOv2-reduced</i>. In the experiment, we use two types of datasets: A <i>SAR ship detection dataset</i> (SSDD) dataset and a <i>Diversified SAR Ship Detection Dataset</i> (DSSDD). These two datasets were used for training and testing purposes. YOLOv2 test results showed an increase in accuracy of ship detection as well as a noticeable reduction in computational time compared to Faster R-CNN. From the experimental results, the proposed YOLOv2 architecture achieves an accuracy of 90.05% and 89.13% on the SSDD and DSSDD datasets respectively. The proposed <i>YOLOv2-reduced</i> architecture has a similarly competent detection performance as YOLOv2, but with less computational time on a NVIDIA TITAN X GPU. The experimental results shows that the deep learning can make a big leap forward in improving the performance of SAR image ship detection.https://www.mdpi.com/2072-4292/11/7/786synthetic aperture radar (SAR) imagesship detectionYOLOv2faster R-CNNYOLOv2-reducedhigh performance computing |
| spellingShingle | Yang-Lang Chang Amare Anagaw Lena Chang Yi Chun Wang Chih-Yu Hsiao Wei-Hong Lee Ship Detection Based on YOLOv2 for SAR Imagery Remote Sensing synthetic aperture radar (SAR) images ship detection YOLOv2 faster R-CNN YOLOv2-reduced high performance computing |
| title | Ship Detection Based on YOLOv2 for SAR Imagery |
| title_full | Ship Detection Based on YOLOv2 for SAR Imagery |
| title_fullStr | Ship Detection Based on YOLOv2 for SAR Imagery |
| title_full_unstemmed | Ship Detection Based on YOLOv2 for SAR Imagery |
| title_short | Ship Detection Based on YOLOv2 for SAR Imagery |
| title_sort | ship detection based on yolov2 for sar imagery |
| topic | synthetic aperture radar (SAR) images ship detection YOLOv2 faster R-CNN YOLOv2-reduced high performance computing |
| url | https://www.mdpi.com/2072-4292/11/7/786 |
| work_keys_str_mv | AT yanglangchang shipdetectionbasedonyolov2forsarimagery AT amareanagaw shipdetectionbasedonyolov2forsarimagery AT lenachang shipdetectionbasedonyolov2forsarimagery AT yichunwang shipdetectionbasedonyolov2forsarimagery AT chihyuhsiao shipdetectionbasedonyolov2forsarimagery AT weihonglee shipdetectionbasedonyolov2forsarimagery |