講演抄録/キーワード |
講演名 |
2017-11-23 11:00
[特別講演]SAR Image Segmentation Based Framework to Ship Detection ○Yang-Lang Chang・Amare Anagaw Ayele(NTUT)・Lena Chang・Wei-Lin Chen(NTOU)・Meng-Che Wu(NSPO)・Chihyuan Chu(NTUT) SANE2017-63 |
抄録 |
(和) |
Synthetic aperture radar (SAR) imagery has proven to be a promising data source for the surveillance of maritime activities, and its application for ship detection has been the focus of many research studies. In this work, we focus on a hierarchical image morphology based approach to perform detection of ship from the given SAR image. The framework does not require any prior knowledge about ships and background observation. The proposed framework has the following four main stages: 1) preprocessing (image alignment, remove unnecessary region and segmentation), 2) sea and land separation, 3) ship detection with feature extraction, and 4) combination of ship information from different segmented image (mark the collected information on the original image). Experiments on real SAR images with varying sea clutter backgrounds and multiple targets situation have been conducted. The effectiveness of the proposed framework is verified using Sentinel-1 data and Terra SAR-X images. The performance analysis confirms that the proposed framework works efficiently in various circumstances with high detection rate of 94.72% in average. |
(英) |
Synthetic aperture radar (SAR) imagery has proven to be a promising data source for the surveillance of maritime activities, and its application for ship detection has been the focus of many research studies. In this work, we focus on a hierarchical image morphology based approach to perform detection of ship from the given SAR image. The framework does not require any prior knowledge about ships and background observation. The proposed framework has the following four main stages: 1) preprocessing (image alignment, remove unnecessary region and segmentation), 2) sea and land separation, 3) ship detection with feature extraction, and 4) combination of ship information from different segmented image (mark the collected information on the original image). Experiments on real SAR images with varying sea clutter backgrounds and multiple targets situation have been conducted. The effectiveness of the proposed framework is verified using Sentinel-1 data and Terra SAR-X images. The performance analysis confirms that the proposed framework works efficiently in various circumstances with high detection rate of 94.72% in average. |
キーワード |
(和) |
feature extraction / ship detection / overlapped based segmentation / hierarchical image segmentation-based framework / / / / |
(英) |
feature extraction / ship detection / overlapped based segmentation / hierarchical image segmentation-based framework / / / / |
文献情報 |
信学技報, vol. 117, no. 321, SANE2017-63, pp. 1-6, 2017年11月. |
資料番号 |
SANE2017-63 |
発行日 |
2017-11-16 (SANE) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
SANE2017-63 |