Rapid detection to long ship wake in synthetic aperture radar satellite imagery

作     者：CHEN Peng LI Xiunan ZHENG Gang 

作者机构：Second Institute of Oceanography Ministry of Natural Resources 

出 版 物：《Journal of Oceanology and Limnology》 (海洋湖沼学报（英文）)

年 卷 期：2019年第37卷第5期

页      面：1523-1532页

核心收录：

学科分类：0710[理学-生物学] 07[理学] 0908[农学-水产] 0707[理学-海洋科学] 0815[工学-水利工程] 

基　　金：Supported by the National Natural Science Foundation of China(No.41476088) the National High Resolution Project of China(No.41Y30B12-9001-14/16) the 2016 Key Projects for Marine Environmental Security(No.2016YFC14032) the research grants of the Second Institute of Oceanography,MNR(No.JT1307) 

主　　题：marine pollution illegal discharge linear fi t oil spill synthetic aperture radar (SAR) long ship wake 

摘      要：The maritime administrative department employs synthetic aperture radar (SAR) satellite remote sensing technology to obtain evidence of illegal discharge of ships. If the ship is discharged during navigation, it forms a long dark wake on the SAR image due to the suppression of the Bragg wave by the oil fi lm. This study investigates key techniques for rapid detection of long ship wakes, thereby providing law enforcement agencies with candidate ships for possible discharge. This paper presents a rapid long ship wake detection method that uses satellite imaging parameters and the axial direction of the ship in images to determine the potential detection area of the wake. Then, the threshold of long ship wake detection is determined using statistical analysis, the area is binarized, and isolated points are removed using a morphological filter operator. The method was tested with ENVISAT Synthetic Aperture Radar and GF-3 SAR data, and results showed that the method was eff ective, and the overall accuracy of the decision reaches 71%. We present two innovations;one is a method that draws a Doppler shift curve, and uses the SAR imaging parameters to determine the detection area of the long wake to achieve rapid detection and reduce the image detection area. The other is where a classical linear fitting method is used to quickly and accurately determine whether the detected dark area is a long ship wake and realizes the twisted long ship wake detection caused by the sea surface flow field, which is otherwise diffi cult to detect by the traditional Radon and Hough transform methods. This method has good suppression performance for the dark spot false alarm formed by low speed wind region or upward flow. The method is developed for maritime ship monitoring system and will promote the operational application of maritime ship monitoring system.