摘要
我国具有漫长的海岸线,专属经济区海域与邻国存在权益争端,必须对他国非法进入我国专属经济区的舰船予以监视监测,维护我国海洋权益。星载SAR具有空间分辨率高、覆盖面积大的特点,能够实现舰船目标的精确定位,且高分辨率SAR图像为舰船类型识别提供了可能,但舰船检测结果难以有效验证。AIS(船舶自动识别系统)能实时进行船舶识别、监视、通信和状态报告,具有定位精度高、全天候的特点,可利用AIS提供的信息对SAR舰船目标检测结果进行有效验证。
目前已有多种基于SAR图像的舰船目标检测算法,但由于缺少实测数据,无法为现有的舰船目标检测结果做出客观的评价。本文开展了星载SAR与AIS舰船目标同步探测实验,获取了青岛附近海域高分辨率RADARSAT-2图像以及同步AIS信息。基于实测数据发展了适用于不同环境下的SAR舰船目标检测方法,开展了高分辨率SAR图像面目标分割技术研究,提高了舰船长宽特征提取精度。
在SAR舰船目标检测算法方面,本文开展不同杂波条件下的舰船目标检测研究,探讨了不同海况、噪声条件对舰船目标检测的影响,提出一种适用于不同杂波条件下的舰船目标检测方法—基于G。分布的SAR舰船目标检测算法,并与经典的基于K分布的CFAR算法对比,利用实测数据对算法进行有效验证。
在高分辨率SAR图像舰船目标分割方面,本文分析多种高分辨率图像滤波算法的滤波效果,并基于高分辨率图像中舰船为面目标的特点,将遥感分类思想引入SAR图像分割领域,建立SAR图像多阈值分割思想,发展了基于FCM的高分辨率SAR影像聚类分割算法,并通过与KSW双阈值算法的分割效果进行对比,评价本文方法的有效性；利用AIS这一重要的辅助数据源,探讨高分辨率SAR影像中舰船目标长、宽信息提取的精度问题。
China has a long coastline, there has Conflicts of interest in the exclusive economic zone with neighboring countries.we must keep watch over the ships entering economic zone Without permission, Maintain our maritime rights and interests. Spaceborne SAR has high spatial resolution and large coverage area, which can determine the precise position of the ship.High-resolution SAR images for the identification of possible ship type, but the ship is difficult to effectively verify the test results. AIS (Automatic Identification System) in real-time identification of ships, surveillance, communications and status reports, with high precision, all-weather features, can be used to provide information on the SAR ship detection results for effective verification.
Nowadays many ship target detection algorithms have been developed based on SAR images. Due to run short of in situ measurement data, it is very difficult to give those algorithms objective evaluation.This paper carried out with the SAR and AIS objective simultaneous detection experiments in the waters near Qingdao, we get high-resolution RADARSAT-2 images and synchronize AIS information. A new ship target detection algorithm has developed based on the measurements,which can apply to different circumstances,and high-resolution SAR image segmentation algorithm of surface targets has developed, which improve the accuracy of ship length and width of feature extraction.
As for SAR Ship target detection algorithm, in this paper ship target detection under different clutter is carried out, the effect of different sea conditions and noise conditions on the ship target detection are discussed, A new algorithm -ship target detection of SAR images based on The G0 distribution which can apply to different circumstances is developed,and with the classical CFAR algorithm based on K distribution in comparison algorithm using measured data on effective verification.
As for high-resolution ship targets segmentation algorithm of SAR images, A variety of high-resolution image filtering algorithm are analyzed.Based on ships in high-resolution images are surface target, remote sensing classification ideas are used in SAR images segmentation,and multi-threshold segmentation theory is established.A high-resolution SAR images clustering segmentation algorithm based on FCM is developed in paper, which were compared with KSW segmentation algorithm to evaluate the effectiveness of this method. AIS are important supplementary data sources,which are used to discuss the accuracy problem of extracting ship target length and width in high resolution SAR images.
引文
[1]Eldhuset K. An automatic ship and ship wake detection system for spaceborne SAR images in coastal regions. IEEE Trans,1996,34(4):1010-1018
[2]Wackerman C C, Friedman K S, Pichel W G, et al. Automatic Detection of Ships in RADARSAT-1 SAR Imagery. Canadian Journal of Remote Sensing,2001,27(5):568-577
[3]Vachon P W, Campbell J W M, Bjerkelund C A, et al. Ship detection by the RADARSAT SAR: validation of detection model predictions. Canadian Journal of Remote Sensing,1997,23(1): 48-59
[4]Vachon P W, Thomas S J, Cranton J, et al. Validation of ship detection by the RADARSAT synthetic aperture radar and the ocean monitoring workstation. Canadian Journal of Remote Sensing,2000,26(3):200-212
[5]种劲松,朱敏慧.SAR图像舰船目标检测算法的对比研究.信号处理,2003,19(6)：580-582
[6]李晓玮,种劲松.基于小波分解的K-分布SAR图像舰船检测.测试技术学报,2007,21(4)：350-354
[7]S.Mallat,A theory for multiresolution signal decomposition:the wavelet representation,IEEE Trans.PAMI,1989,11(7),674-693.
[8]Leonid I.Perlovsky, William H.Svhoendorf,Model base neural network for target detection in SAR images,IEEE.Trans.on Image Processing,1997,6(1),203-215.
[9]Kraut Eldhuset,An automatic ship and ship wake detection system for spaceborne SAR images in coastal regions,IEEE Trans.on,G.R.S,1996 34(4),1010-1019.
[10]Quoc H pham,Timothy M.Brosnan,Multistage Algorithm for detection target in SAR image date,SPIE,1997,3070,66-75.
[11]L.Benyoussef,Y.Delignon,An optimal matched filter for target detection in images distorted by noise,IGARSS'1998,1000-1003.
[12]Li Y,Gong P,Classification and Segmentation of Remote Sensing Images[J].International Journal of Remote Sensing,in press.
[13]Yu Q,Gong P,Clinton N,et al.Object-based Detailed Vagetation Mapping Using High Spatial Resolution Imagery[J].Photogrammetric Engineering and Remote Sensing,in press
[14]Li X,Yeh A G.Multitemporal SAR Images for Monitoring Cultivation Systems Using Case-based Reasoning[J]. Remote Sensing of Environment,2004,90(4):524-534.
[15]Kelly P A, Derin H. Adaptive segmentation of speckled images using hierachical random field model[J], IEEE Trans. On ASSP,1988,36(10):1628-1641.
[16]Ives R W, Eichel P. Application of pixel segmentation to thelow rate compression of complex SAR imagery,IGARSS'98,1998:1064-1067.
[17]Derin H, Kelly P,Modeling and segmentation of speckled images using complex data[J]. IEEE Trans. Geosci. Remote Sensing,1990,28(1):76-87.
[18]付琨,匡纲要,郁文贤.高分辨率SAR图像地物分类算法研究[J].电子学报,2001,29(12A)：1820-1823.
[19]Bachmann C M, Musman S A Schultz A. Lateral Inhibition Neural Networks for Classification of Simulated Radar Imagery. IEEE International Conference on Neural Networks, Baltimore,1992,2:115-120
[20]Osman H, Blostein S, Gagnon L. Classification of Ships in Airborne SAR Imagery Using Backpropagation Neural Networks. SPIE Proc.1997,3161:126-136
[21]Askari F, Zerr B. Automatic approach to ship detection in spaceborne synthetic aperture radar imagery:An assessment of ship detection capability using RADARSAT. Technical Report SACLANTCEN-SR-338, Italy:SACLANT Undersea Research Centre, La Spezia,2000.36
[22]林长川.雷达与AIS目标位置信息融合方法的研究[J].中国航海.2002(1)：22-25.
[23]迟卫,石爱国,张永胜等.军船自动识别系统研究[J].舰船科学技术,2003(10)：66-68.
[24]林长川.雷达与AIS.目标航迹模糊关联与统计加权合并融合方法的探讨[J].中国航海,2003(4)：70-73.
[25]种劲松,欧阳越,朱慧敏.合成孔径雷达图像海洋目标检测.北京：海洋出版社,2006.2-105
[26]Hughes, Effect of the Kelvin Wake Transverse Wave Component on SAR Imagery of Surface Wake Wakes, Proceedings of IGARSS'89 Symposium, Vancouver, Canada,10-14 July,1989
[27]纪永刚,张杰,张晰.基于现场同步实验的星载SAR舰船目标尺度估计分析.首届海洋遥感技术交流会论文集,广西北海,2007.
[28]Gordon C. Staples, Ship Detection using RADARSAT SAR Imagery, Geomatics in the Era of RADARSAT, May 25-30,1997, Ottawa, Canada
[29]Floyd M. Hendeeson and Anthony J.Lewis, Principles&Applications of Imaging Radar,Manual of Remote Sensing,Third Edition, Volume 2, John Wiley&Sons, Inc.
[30]阮颖铮等,《雷达截面与隐身技术》,国防工业出版社,2001.
[31]M. Skolnik, An Empirical Formula for the Radar Cross Section of Ships at Grazing lncidence,IEEE Trans. on Aerospace and Electronic System, Vol.10, March 1974.
[32]M. Skolnik, Introduction to Radar Systems, Second Edition, MeGraw-Hill Baok Company,1980
[33]SHAO M, NIKIAS C L. Signal processing with fractional lower-order moments:Stable processes and their applications[J].Proc.IEEE,1993,81(7):986-1010.
[34]JIANG Q S, etal.Automatic detection for ship targets in RADARSAT SAR images from coastal regions[J].Vision Interface'99,Trois-Tividres,Canada,May 1999,131-137.
[35]FRERY A C, etal.A model for extremely heterogeneous clutter[J].IEEE Trans.On GRS,1997,35(3):648-659.
[36]贺志国,周晓光,陆军等.一种基于G0分布的SAR图像快速CFAR检测方法[J].国防科技大学学报,2009,31(1)：47-51.
[37]C.J. Oliver. Understanding Synthetic Aperture Radar Images. Artech House, Boston/London,1998.
[38]高贵,鲁敏,黄纪军,匡纲要,李德仁.高分辨SAR图像中杂波的统计特性分析.信号处理,2008,24(4)：648-654.
[39]OSCAR H, LUCINI M M, FRERY A C. M-Estimators of Roughness and Scale for GAO-Modelled SAR Imagery[J].EURASIP Journal on Applied Signal Processing,2002(1):105-114.
[40]时公涛,高贵,周晓光等.基于Mellin变换的G0分布参数估计方法[J].自然科学进展,2009,19(6)：677-690.
[41]MAITRE H合成孔径雷达图像处理[M].孙洪,译.北京：电子工业出版社,2005：67-107.
[42]OLIVER C J.Understanding Synthetic Aperture Radar Images[M].Artech House,Boston/London,1998.
[43]黄树生.南麂海域海况特征分析[J].海洋通报,1998,17(3)：90-96.
[44]张晰.星载SAR舰船目标探测实验研究：[硕士学位论文].青岛：中国海洋大学,2007.
[45]苏创业.SAR图像斑点噪声抑制算法研究与应用：[硕士学位论文].厦门：厦门大学,2008.
[46]NOEL R, et al. Ship Surveillance using RADARSAT ScanSAR images[C],Ship Detection in Coastal Waters Workshop 2000,NS,Canada,31M ay,1 & 2 June,2000:41-45.
[47]NOVAK L M, SECHEN M B, CARDULLO M J.Studies of target detection algorithms that use polarimetric data[J].IEEE Trans.on Aerospace and Electronic Systems,1989,25(2):150-165.
[48]种劲松.合成孔径雷达图像舰船目标检测算法与应用研究：[博士学位论文].北京：中国科学院电子学研究所,2002.
[49]S. Noushath, G. H. Kumar, and P. Shivakumara. Diagonal fisher linear discriminant analysis for efficient face recognition. Neurocomputing, Vol.69,2006:1711-1716.
[50]A. Likas, N. Vlassis and J. J. Verbeek. The global k-means clustering algorithm.Pattern Recognition, Vol.36,2003:451-461.
[51]胡利平.合成孔径雷达图像目标识别技术研究：[博士学位论文].西安：西安电子科技大学,2009.
[52]刘永俊,陈才扣.最大散度差鉴别分析及人脸识别.计算机工程与应用,Vol.34,2006：208-210.
[53]Song F X, Zhang D, Mei D Y, et al. A multiple maximum scatter difference discriminant criterion for facial feature extraction. IEEE Trans. on System, Man,and Cybernetic-Part B Cybernetics, Vol.37(6),,2007:1599-1606
[54]Kuan D T,Sawchuk A A,Strand T C,etal A daptive restoration of in ages with speckle[J].IEEE Transactions on A coustics,Speech and Singal Processing 1987,35(3):373-383.
[55]Lopes A,Nezry E,Touzi R,et al Maxinum a posteriori filtering and first order texture models in SAR images[A].In Proceedings of IEEE International Geoscience and Remote Sensing Symposium[C], Washington D C,1990,2409-2412.
[56]P. V. Narasimha Rao.An Adaptive Filter for Speckle.Suppression in synthetic Aperture Radar Images.INT.J.Remote Sensing,1995,16(5),877-889.
[57]J. C.Dunn. A fuzzy relative of the ISODATA process and its use in detecting compact, well-separated clustern. J.Cybern,1974.3;32-57
[58]J. C. Bezdek. Pattern recognition with fuzzy objective function algorithms. Plenum,New Yurk.1981
[59]R. L. Cannon.J. Dave and J. C. Bezdek. Efficient implementation of the fussy c-means clustering algorithms.IEEE Trans,1986;PAMI-8(2):248-255
[60]Kapur J N., Sahoo I'K., along A K C, A New Method of Gray-level Picture Thresholding Using the Entropy of the Histagram, Computer Vision, Graphics, and Image Processing,1985 Vol.29, No.2,273-285
[61]陈述彭,遥感大词典,北京：科学出版社,1990