高分辨率SAR仿真图像生成技术及其在目标识别中的应用
|
摘要
合成孔径雷达(Synthetic Aperture Radar,SAR)在地面目标识别中有着广泛的应用需求。SAR图像会随着观测视角、雷达带宽等变化而变化,为实现高识别率,通常需要大量的不同成像条件下SAR图像。由于客观条件限制,这一要求通常难以满足。利用仿真数据实现高分辨率SAR图像生成是雷达目标识别研究中重要的数据源产生手段,因此,研究高分辨率SAR图像生成技术具有重要的现实意义。
本文以SAR图像识别需求为牵引,对SAR图像仿真的实现方法进行了大量的研究,在此基础上,研究了仿真图像在目标识别上的应用,具体来说包括以下三个方面:
(1)研究基于转台成像原理的仿真SAR图像生成方法。分析转台成像原理及其在SAR图像仿真上的一致性,针对方位向大转角和小转角两种情况下回波数据的特点,分析并实现了不同的处理方法,利用电磁散射计算软件RadBase得到目标的电磁散射数据,得到了高分辨率SAR仿真图像。
(2)对广义相关法进行了改进,提出了一种能有效去除遮挡效应的改进广义相干SAR仿真图像生成方法,解决了SAR图像仿真中的遮掩问题。SAR图像仿真的意义在于能够提供目标识别的数据源,但是遮挡效应使得SAR图像敏感与目标姿态,从而需要目标识别模板库较为密集。本文以不同观测视角下的大转角成像为依据,采用一种非相干的处理方法来去除遮挡效应,使得模板库内的图像较为完整。利用这一方式可有效降低图像随方位角变化特性,提高建库效率。
(3)基于仿真数据相关性分析,提出了一种识别模板库的设计思路。仿真SAR图像可用于建立SAR图像数据库,支持目标识别技术研究,本文首先分析了仿真图像用于图像检测和目标识别的方法;在此基础上,对SAR图像模板库的建库规模进行了研究,提出了最优模板数据库的仿真设计方法。
论文针对基于SAR图像的目标识别中SAR图像生成进行了详细的研究,在地面目标尤其是地面军事目标识别中具有较强的现实意义,希望能够为未来的高分辨星载SAR系统在地面目标识别中的应用服务,为目标识别数据库的建立提供帮助。
The Synthetic Aperture Radar (SAR) has broad application in ground target recognition. In normal, the SAR image varying with the observation angle or radar band width, as a result large number of SAR imaging under different condition was used for better identification. Due to the constraints of objective condition, acquisitions of so many SAR images is impossible, Simulated SAR image is one of the most important data source for target recognition. Thereby study of SAR image simulating under different condition have important realistic meaning.
Based on the needing of SAR image identification, many simulating methods for SAR image were studied in this paper, and then the simulated images were applied in target recognition. The main content of this paper was concluded as following:
(1) The paper studied the simulated SAR image produce technology based on rotating target mode. Firstly, the consistency between SAR imaging and rotate target imaging is analyzed. According to the characteristic of narrow angle mode and wide angle mode, different processing algorithm is analyzed. The data is simulated based on software Radbase. And the simulated SAR image is got under this scheme.
(2) The paper modified the generalized coherent imaging algorithm. A imaging algorithm which can avoid the shading effect is proposed. And the problem of shading is solved by this way. The meaning of simulating SAR image is to supply data source for target recognition. But shading effect made the target SAR image is sensitive to the look angle. The paper used the modified coherent process to get the complete SAR image. Thus the image in the template base will be relative intact. So the sensitive to the angle can be reduced. The efficiency to create database will be improved.
(3) Base on simulated SAR images coherence, a template database design scheme is proposed. The simulated SAR image could construct SAR image database, which was utilized for target recognition. First, the method about image detection and target recognition by simulated image was analyzed, based on which the scale of SAR image template database was studied, and the simulated method for optimal template database was proposed.
In this paper, the SAR image generating method was studied in detail, which is important for target recognition based on SAR images. The study result is expected to serve in application of ground target recognition in future high resolution space borne SAR system, at least be of help for target recognition database construction.
本文以SAR图像识别需求为牵引,对SAR图像仿真的实现方法进行了大量的研究,在此基础上,研究了仿真图像在目标识别上的应用,具体来说包括以下三个方面:
(1)研究基于转台成像原理的仿真SAR图像生成方法。分析转台成像原理及其在SAR图像仿真上的一致性,针对方位向大转角和小转角两种情况下回波数据的特点,分析并实现了不同的处理方法,利用电磁散射计算软件RadBase得到目标的电磁散射数据,得到了高分辨率SAR仿真图像。
(2)对广义相关法进行了改进,提出了一种能有效去除遮挡效应的改进广义相干SAR仿真图像生成方法,解决了SAR图像仿真中的遮掩问题。SAR图像仿真的意义在于能够提供目标识别的数据源,但是遮挡效应使得SAR图像敏感与目标姿态,从而需要目标识别模板库较为密集。本文以不同观测视角下的大转角成像为依据,采用一种非相干的处理方法来去除遮挡效应,使得模板库内的图像较为完整。利用这一方式可有效降低图像随方位角变化特性,提高建库效率。
(3)基于仿真数据相关性分析,提出了一种识别模板库的设计思路。仿真SAR图像可用于建立SAR图像数据库,支持目标识别技术研究,本文首先分析了仿真图像用于图像检测和目标识别的方法;在此基础上,对SAR图像模板库的建库规模进行了研究,提出了最优模板数据库的仿真设计方法。
论文针对基于SAR图像的目标识别中SAR图像生成进行了详细的研究,在地面目标尤其是地面军事目标识别中具有较强的现实意义,希望能够为未来的高分辨星载SAR系统在地面目标识别中的应用服务,为目标识别数据库的建立提供帮助。
The Synthetic Aperture Radar (SAR) has broad application in ground target recognition. In normal, the SAR image varying with the observation angle or radar band width, as a result large number of SAR imaging under different condition was used for better identification. Due to the constraints of objective condition, acquisitions of so many SAR images is impossible, Simulated SAR image is one of the most important data source for target recognition. Thereby study of SAR image simulating under different condition have important realistic meaning.
Based on the needing of SAR image identification, many simulating methods for SAR image were studied in this paper, and then the simulated images were applied in target recognition. The main content of this paper was concluded as following:
(1) The paper studied the simulated SAR image produce technology based on rotating target mode. Firstly, the consistency between SAR imaging and rotate target imaging is analyzed. According to the characteristic of narrow angle mode and wide angle mode, different processing algorithm is analyzed. The data is simulated based on software Radbase. And the simulated SAR image is got under this scheme.
(2) The paper modified the generalized coherent imaging algorithm. A imaging algorithm which can avoid the shading effect is proposed. And the problem of shading is solved by this way. The meaning of simulating SAR image is to supply data source for target recognition. But shading effect made the target SAR image is sensitive to the look angle. The paper used the modified coherent process to get the complete SAR image. Thus the image in the template base will be relative intact. So the sensitive to the angle can be reduced. The efficiency to create database will be improved.
(3) Base on simulated SAR images coherence, a template database design scheme is proposed. The simulated SAR image could construct SAR image database, which was utilized for target recognition. First, the method about image detection and target recognition by simulated image was analyzed, based on which the scale of SAR image template database was studied, and the simulated method for optimal template database was proposed.
In this paper, the SAR image generating method was studied in detail, which is important for target recognition based on SAR images. The study result is expected to serve in application of ground target recognition in future high resolution space borne SAR system, at least be of help for target recognition database construction.
引文
[1] European Space Agency. Advanced Synthetic Aperture Radar. http://envisat. esa.int/instruments/asar/. 2007-3-6.
[2] Shimada M. JERS-1 SAR User hand book. http://www. eorc .jaxa.jp/JERS一1/user-handbook/ index.html. 2003-4-6.
[3] The German Aerospace Center (DLR). TerraSAR-X Mission. http://www. dlr.de/tsx/main/ mission en.html. 2008-3-29.
[4] MacDonald D，Associates Ltd. Radarsat-2 Mission. http://www. radarsat2.info/about/mission. asp. 2008-2-5.
[5] Italian Space Agency. General Information on COSMO-SkyMed System, Products and AO. http://cosmo-skymed-ao.asi.it/asi/asi. 2008-2-5.
[6]国家减灾委.环境与灾害监测预报小卫星星座项目概况、组成及特点. http://www.jianzai. gov.cn/index.html. 2008-2-5.
[7]郭华东.汉川地震灾害遥感图集.北京:科学出版社.2008.
[8]熊谦，郭峰.FFT在ISAR成像中的应用[J].测控技术，2004, 23(10):15-16
[9]阮颖铮等著.雷达截面与隐身技术[M].北京:国防工业出版，1998.
[10] J.L.Walker, Range-Doppler Imaging of Rotating Objects, IEEE Trans, on AES, 1980,16(1): 23-52.
[11] S.M.Kay, Modern spectral estimation theory and application. Englewood Cliffs, NJ: Prentice-Hall, 1988.
[12] D.A.Ausherman, A.Kozma, J.L.Walker et al. Developments in Radar Imaging. IEEE Trans. On Aerospace and Electronic Systems, Vol.AES-8,No.2March 1972.
[13] Masayoshi Tsuchida, Kei Suwa, and Masafumi Iwamoto, A simulator of synthetic aperture radar images; land ocean surface, and man-made target, IGARSS 2004,4836-4838.
[14] Rajan Bhalla, Hao Lee Ling, Shung Wu, Jeff. Hughes. Bistatic Scatering Center Extraction Using the Shooting and Bouncing Ray Technique. Proceedings of SPIE. The International Society for Optical Engineering.1999,(3721):612-619.
[15]匡纲要，技科峰，粟毅，郁文贤.SAR图像自动目标识别研究.中国图像图形学报，Vol.B(A):1115-1120, 2003.
[16]袁礼海，宋建社，薛文通，赵伟舟.SAR图像自动目标识别系统研究与设计[J] .计算机应用研究，Vol. 11, 249-251, 2006.
[17] T. Ross, S. Worrell, V. Velten, J. Mossing, and M. Bryant. Standard SAR ATR evaluation experiments using the MSTAR public release data set. Part of the SPIE Conference on Algorithms for Synthetic Aperture Radar Imagery V, Vo1.3370, 566-573, 1998.
[18] M. Bryant, S. Worrell, and A. Dixon. MSE template size analysis for MSTAR data. Part of the SPIE Conference on Algorithms for Sythetic Aperture Radar Imagery V, Vol. 3370,396-405,1998.
[19] B. Bhanu. Automatic target recognition: state of the art survey. IEEE Trans. on Aerospace and Electronic Systems,1986,Vo1. 22(4), 364-379.
[20] J. A. Ratches,Wapes, et al.Aided and automatic target recognition based upon sensory inputs from images forming systems. IEEE Trans. on Pattern Analysis and Machine Intelligence 1997，Vol. 19(9), 1004-1019.
[21] Q. Zhao and J. C. Principe. Support vector machines for SAR automatic target recognition. IEEE Trans. on Aerospace and Electronic Systems，VO1.37(2)，643-654，2001.
[22]庄钊文，袁乃昌编著.雷达散射截面测量一紧凑场理论与技术[D].国防科技大学出版社.2000.
[23]黄培康，殷红成，许小剑.雷达目标特性[M].北京:电子工业出版社，2005.
[24]钟顺时，钮茂德.电磁场理论基础[M].西安电子科技大学出版社，1995.
[25]胡广书编著.数字信号处理一理论、算法与实现[M].北京:清华大学出版社. 1997.
[26] D.Pastina，A.Farina，FIEE，J.Gunning，P.Lombardo. Two-dimensional super-resolution spectral analysis applied to SAR images, IEE Pro-Radar Sonar Navig. Vo1.145，No.S，October 1998:281-290.
[27] Stuart R.DeGraaf. SAR Imaging via Modern 2-D Spectral Estimation Methods. IEEE Trans. On Image Processing. Vol.4 , 1998:681-703.
[28]保铮，邢孟道，王彤.雷达成像技术[M].北京:电子工业出版社，2005.
[29] RIUS J M，FERRANDOM, JOFRE L. High-frequency RCS of complex radar targets in real-time[J].IEEE Transactions on Antennas and Propagation Vo1.35，No.2，1993:1308-1319.
[30]庄钊文等著.军用目标雷达散射界面预估与测量[M].北京:科学出版社，2007.
[31]孙即祥等.模式识别中的特征提取与计算机视觉不变量.北京:国防工业出版社.2001.
[32]丁纬，朱兆达.ECP成像的子图像叠加法.信号处理，Vol.13' No.4' 1997: 312-317.
[33]丁纬，朱兆达，叶泰如.对广义相干处理成像算法的改进.数据采集与处理，Vol.ll，No.2，1996: 107-111.
[34]夏听.SAR图像典型目标特征提取与识别方法研究.四川大学硕士学问论文，2006.
[35]计科峰，匡纲要，郁文贤.基于线性回归的SAR目标方位角估计方法2004,26(11): 26-29.
[36]张翠，丽fi苏丹，王正志.SAR图像目标分割与方位角估计.国防科技大学学报，2001，23(5): 74-78.
[37] Tupin F ' Maitre H ' Mangin J F et al·Detection of linear features in SAR images. application road network extraction[J].IEEE Trans. Geoscience and Remote Sensing, 1998，36(2):434-453.
[38] Ian G. Gumming, Frank H. Wong, Digital Processing of Synthetic Aperture Radar Data: Algorithms and Implementation [M]. 2004.
[39]刘永坦等.雷达成像技术[M].哈尔滨工业大学出版社.1999.
[40]保铮，邢孟道，王彤.雷达成像技术[M].北京:电子工业出版社.2005.
[41] C.V. Jakowatz, D.E. Wahl and P.H. Eichel，Spotlight-Mode Synthetic Aperture Radar: A Signal Processing Approach Kluwer Academic Publishers Boston MA} 1996.
[42] M.R.Allen and L.E.Holl，“Wide-angle wideband SAR matched filter image formation for enhanced detection performance,”in Algorithms for Synthetic Aperture Radar Imagery , 2230:302-314 , SPIE . (Orlando FL), April 1994.
[43]金光虎.中段弹道目标ISAR成像及物理特性反演技术研究[D].国防科学技术大学博士论文.2009.
[44]张红，王超，张波等.高分辨率SAR图像目标识别[M].北京:科学出版社.2009.
[45] Moreira A. Real-Time Synthetic Aperture Radar Processing with a New Sub- aperture Approach, IEEE Trans. on GRS, 30(4):714-722. 7u1. 1992.
[46]孙即祥，等编著.现代模式识别[M].国防科技大学出版社.2002.
[2] Shimada M. JERS-1 SAR User hand book. http://www. eorc .jaxa.jp/JERS一1/user-handbook/ index.html. 2003-4-6.
[3] The German Aerospace Center (DLR). TerraSAR-X Mission. http://www. dlr.de/tsx/main/ mission en.html. 2008-3-29.
[4] MacDonald D，Associates Ltd. Radarsat-2 Mission. http://www. radarsat2.info/about/mission. asp. 2008-2-5.
[5] Italian Space Agency. General Information on COSMO-SkyMed System, Products and AO. http://cosmo-skymed-ao.asi.it/asi/asi. 2008-2-5.
[6]国家减灾委.环境与灾害监测预报小卫星星座项目概况、组成及特点. http://www.jianzai. gov.cn/index.html. 2008-2-5.
[7]郭华东.汉川地震灾害遥感图集.北京:科学出版社.2008.
[8]熊谦，郭峰.FFT在ISAR成像中的应用[J].测控技术，2004, 23(10):15-16
[9]阮颖铮等著.雷达截面与隐身技术[M].北京:国防工业出版，1998.
[10] J.L.Walker, Range-Doppler Imaging of Rotating Objects, IEEE Trans, on AES, 1980,16(1): 23-52.
[11] S.M.Kay, Modern spectral estimation theory and application. Englewood Cliffs, NJ: Prentice-Hall, 1988.
[12] D.A.Ausherman, A.Kozma, J.L.Walker et al. Developments in Radar Imaging. IEEE Trans. On Aerospace and Electronic Systems, Vol.AES-8,No.2March 1972.
[13] Masayoshi Tsuchida, Kei Suwa, and Masafumi Iwamoto, A simulator of synthetic aperture radar images; land ocean surface, and man-made target, IGARSS 2004,4836-4838.
[14] Rajan Bhalla, Hao Lee Ling, Shung Wu, Jeff. Hughes. Bistatic Scatering Center Extraction Using the Shooting and Bouncing Ray Technique. Proceedings of SPIE. The International Society for Optical Engineering.1999,(3721):612-619.
[15]匡纲要，技科峰，粟毅，郁文贤.SAR图像自动目标识别研究.中国图像图形学报，Vol.B(A):1115-1120, 2003.
[16]袁礼海，宋建社，薛文通，赵伟舟.SAR图像自动目标识别系统研究与设计[J] .计算机应用研究，Vol. 11, 249-251, 2006.
[17] T. Ross, S. Worrell, V. Velten, J. Mossing, and M. Bryant. Standard SAR ATR evaluation experiments using the MSTAR public release data set. Part of the SPIE Conference on Algorithms for Synthetic Aperture Radar Imagery V, Vo1.3370, 566-573, 1998.
[18] M. Bryant, S. Worrell, and A. Dixon. MSE template size analysis for MSTAR data. Part of the SPIE Conference on Algorithms for Sythetic Aperture Radar Imagery V, Vol. 3370,396-405,1998.
[19] B. Bhanu. Automatic target recognition: state of the art survey. IEEE Trans. on Aerospace and Electronic Systems,1986,Vo1. 22(4), 364-379.
[20] J. A. Ratches,Wapes, et al.Aided and automatic target recognition based upon sensory inputs from images forming systems. IEEE Trans. on Pattern Analysis and Machine Intelligence 1997，Vol. 19(9), 1004-1019.
[21] Q. Zhao and J. C. Principe. Support vector machines for SAR automatic target recognition. IEEE Trans. on Aerospace and Electronic Systems，VO1.37(2)，643-654，2001.
[22]庄钊文，袁乃昌编著.雷达散射截面测量一紧凑场理论与技术[D].国防科技大学出版社.2000.
[23]黄培康，殷红成，许小剑.雷达目标特性[M].北京:电子工业出版社，2005.
[24]钟顺时，钮茂德.电磁场理论基础[M].西安电子科技大学出版社，1995.
[25]胡广书编著.数字信号处理一理论、算法与实现[M].北京:清华大学出版社. 1997.
[26] D.Pastina，A.Farina，FIEE，J.Gunning，P.Lombardo. Two-dimensional super-resolution spectral analysis applied to SAR images, IEE Pro-Radar Sonar Navig. Vo1.145，No.S，October 1998:281-290.
[27] Stuart R.DeGraaf. SAR Imaging via Modern 2-D Spectral Estimation Methods. IEEE Trans. On Image Processing. Vol.4 , 1998:681-703.
[28]保铮，邢孟道，王彤.雷达成像技术[M].北京:电子工业出版社，2005.
[29] RIUS J M，FERRANDOM, JOFRE L. High-frequency RCS of complex radar targets in real-time[J].IEEE Transactions on Antennas and Propagation Vo1.35，No.2，1993:1308-1319.
[30]庄钊文等著.军用目标雷达散射界面预估与测量[M].北京:科学出版社，2007.
[31]孙即祥等.模式识别中的特征提取与计算机视觉不变量.北京:国防工业出版社.2001.
[32]丁纬，朱兆达.ECP成像的子图像叠加法.信号处理，Vol.13' No.4' 1997: 312-317.
[33]丁纬，朱兆达，叶泰如.对广义相干处理成像算法的改进.数据采集与处理，Vol.ll，No.2，1996: 107-111.
[34]夏听.SAR图像典型目标特征提取与识别方法研究.四川大学硕士学问论文，2006.
[35]计科峰，匡纲要，郁文贤.基于线性回归的SAR目标方位角估计方法2004,26(11): 26-29.
[36]张翠，丽fi苏丹，王正志.SAR图像目标分割与方位角估计.国防科技大学学报，2001，23(5): 74-78.
[37] Tupin F ' Maitre H ' Mangin J F et al·Detection of linear features in SAR images. application road network extraction[J].IEEE Trans. Geoscience and Remote Sensing, 1998，36(2):434-453.
[38] Ian G. Gumming, Frank H. Wong, Digital Processing of Synthetic Aperture Radar Data: Algorithms and Implementation [M]. 2004.
[39]刘永坦等.雷达成像技术[M].哈尔滨工业大学出版社.1999.
[40]保铮，邢孟道，王彤.雷达成像技术[M].北京:电子工业出版社.2005.
[41] C.V. Jakowatz, D.E. Wahl and P.H. Eichel，Spotlight-Mode Synthetic Aperture Radar: A Signal Processing Approach Kluwer Academic Publishers Boston MA} 1996.
[42] M.R.Allen and L.E.Holl，“Wide-angle wideband SAR matched filter image formation for enhanced detection performance,”in Algorithms for Synthetic Aperture Radar Imagery , 2230:302-314 , SPIE . (Orlando FL), April 1994.
[43]金光虎.中段弹道目标ISAR成像及物理特性反演技术研究[D].国防科学技术大学博士论文.2009.
[44]张红，王超，张波等.高分辨率SAR图像目标识别[M].北京:科学出版社.2009.
[45] Moreira A. Real-Time Synthetic Aperture Radar Processing with a New Sub- aperture Approach, IEEE Trans. on GRS, 30(4):714-722. 7u1. 1992.
[46]孙即祥，等编著.现代模式识别[M].国防科技大学出版社.2002.