English  |  正體中文  |  简体中文  |  Items with full text/Total items : 26988/38789
Visitors : 2324722      Online Users : 62
RC Version 4.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Adv. Search
LoginUploadHelpAboutAdminister

Please use this identifier to cite or link to this item: http://ntour.ntou.edu.tw:8080/ir/handle/987654321/52270

Title: Ship detection based on YOLOv2 for SAR imagery
Authors: Yang-Lang Chang
Amare Anagaw
Lena Chang
Yi Chun Wang
Chih-Yu Hsiao
Wei-Hong Lee
Contributors: 國立臺灣海洋大學:通訊與導航工程學系
Keywords: synthetic aperture radar (SAR) images
ship detection
YOLOv2
faster R-CNN
YOLOv2-reduced
high performance computing
Date: 2019-03
Issue Date: 2019-06-21T02:29:58Z
Publisher: Remote Sensing
Abstract: Abstract: Synthetic aperture radar (SAR) imagery has been used as a promising data source formonitoring maritime activities, and its application for oil and ship detection has been the focus ofmany previous research studies. Many object detection methods ranging from traditional to deeplearning approaches have been proposed. However, majority of them are computationally intensiveand have accuracy problems. The huge volume of the remote sensing data also brings a challengefor real time object detection. To mitigate this problem a high performance computing (HPC) methodhas 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 theSAR images. The You Only Look Once version 2 (YOLOv2) deep learning framework is proposedto model the architecture and training the model. YOLOv2 is a state-of-the-art real-time objectdetection system, which outperforms Faster Region-Based Convolutional Network (Faster R-CNN) andSingle Shot Multibox Detector (SSD) methods. Additionally, in order to reduce computational timewith relatively competitive detection accuracy, we develop a new architecture with less number oflayers called YOLOv2-reduced. In the experiment, we use two types of datasets: A SAR ship detectiondataset (SSDD) dataset and a Diversified SAR Ship Detection Dataset (DSSDD). These two datasets wereused for training and testing purposes. YOLOv2 test results showed an increase in accuracy of shipdetection as well as a noticeable reduction in computational time compared to Faster R-CNN. Fromthe experimental results, the proposed YOLOv2 architecture achieves an accuracy of 90.05% and89.13% on the SSDD and DSSDD datasets respectively. The proposed YOLOv2-reduced architecturehas a similarly competent detection performance as YOLOv2, but with less computational time on aNVIDIA TITAN X GPU. The experimental results shows that the deep learning can make a big leapforward in improving the performance of SAR image ship detection.
Relation: 11(7) pp.786
URI: http://ntour.ntou.edu.tw:8080/ir/handle/987654321/52270
Appears in Collections:[通訊與導航工程學系] 期刊論文

Files in This Item:

File Description SizeFormat
index.html0KbHTML2View/Open


All items in NTOUR are protected by copyright, with all rights reserved.

 


著作權政策宣告: 本網站之內容為國立臺灣海洋大學所收錄之機構典藏,無償提供學術研究與公眾教育等公益性使用,請合理使用本網站之內容,以尊重著作權人之權益。
網站維護: 海大圖資處 圖書系統組
DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback