空间目标特征提取及识别技术
|
摘要
当前空间技术迅猛发展,现代战争对空间依赖性增强,人类航天活动增加,空间目标监视和识别变得尤为重要。现役的空间目标特性测量雷达大多属于低分辨窄带雷达,虽然其信息量有限,但造价相对低廉,目前仍是一种主要的空间目标探测传感器。对窄带雷达来说,虽然其距离分辨率较低,但受不同目标散射特性的影响,不同目标间的回波包络会有一定差异,这一差异信息可以用于实现目标识别。对空间目标,由于其运动有自身的规律,不像飞机等目标相对于雷达的姿态变化具有较大不确定性。因此,采用窄带雷达对空间目标进行识别具有较大实用价值,这也是本文研究的出发点。
本文研究了空间目标的雷达回波特性和识别方法,分析了窄带雷达实测数据的分布情况,然后对其提取具有平移不变性的中心矩特征。再分别用主成分分析法(PCA)和Fisher线性判别法(FLD)对中心矩特征进一步选择,最后分别用最小欧氏距离法和支持向量机(SVM)对空间目标分类识别。通过实测数据的实验结果验证了利用雷达回波包络差异识别空间目标是可行的。
Currently, rapid and drastic changes are taking place in the space technology. With the reinforcement of depending on space in modern wars and the development of spaceflight activity of human, the surveillance and recognition of space targets become especially important. Most of active radars, used to measure target's characteristic, belong to low-resolution narrowband radar. The information obtained by low-resolution narrowband is limited, but it was still a main sort of space target detect sensor for low cost. The range resolution of narrowband radar is low. Influenced by different target scattering characteristics, there are differences in echo envelop of different targets, which could be used for target recognition. The movement of space target is regular, so variation of its pose is certain. The research starting point of this thesis lies that satisfactory results can achieved by using narrowband radar to recognize space target.
In this thesis, the radar echo characteristic and the identification of space target are studied, and the distribution of measured data is analyzed. Their central moment's features, with translation-invariant property, are calculated for further recognition. Then Principle Component Analysis (PCA) and Fisher Linear Discriminant (FLD) are respectively used to choose the central moment feature. Finally, the Least Euclidean Distance and a multi-class support vector machine (SVM) classifier are respectively designed to classify space targets. The experimental comparisons based on measured satellites data show that the space target can be recognized by utilizing its radar echo envelop.
本文研究了空间目标的雷达回波特性和识别方法,分析了窄带雷达实测数据的分布情况,然后对其提取具有平移不变性的中心矩特征。再分别用主成分分析法(PCA)和Fisher线性判别法(FLD)对中心矩特征进一步选择,最后分别用最小欧氏距离法和支持向量机(SVM)对空间目标分类识别。通过实测数据的实验结果验证了利用雷达回波包络差异识别空间目标是可行的。
Currently, rapid and drastic changes are taking place in the space technology. With the reinforcement of depending on space in modern wars and the development of spaceflight activity of human, the surveillance and recognition of space targets become especially important. Most of active radars, used to measure target's characteristic, belong to low-resolution narrowband radar. The information obtained by low-resolution narrowband is limited, but it was still a main sort of space target detect sensor for low cost. The range resolution of narrowband radar is low. Influenced by different target scattering characteristics, there are differences in echo envelop of different targets, which could be used for target recognition. The movement of space target is regular, so variation of its pose is certain. The research starting point of this thesis lies that satisfactory results can achieved by using narrowband radar to recognize space target.
In this thesis, the radar echo characteristic and the identification of space target are studied, and the distribution of measured data is analyzed. Their central moment's features, with translation-invariant property, are calculated for further recognition. Then Principle Component Analysis (PCA) and Fisher Linear Discriminant (FLD) are respectively used to choose the central moment feature. Finally, the Least Euclidean Distance and a multi-class support vector machine (SVM) classifier are respectively designed to classify space targets. The experimental comparisons based on measured satellites data show that the space target can be recognized by utilizing its radar echo envelop.
引文
[1]马君国,王远模,李保国等.基于距离像序列的空间目标识别算法.现代雷达.2006,4.28(4)：25-28
[2]李海顺,董天临,曹毅强.空间测量雷达目标分类与识别算法及其能力分析.舰船电子工程.2005,3,25(3)：106-109.
[3]李玉书.空间目标监视雷达技术探讨.飞行器测控学报.2003.22(4)：62-66.
[4]魏晨曦,俄罗斯的空间目标监视、识别、探测与跟踪系统.中国航天.2006,8：39-41.
[5]刘绍球,高淑霞,高永浩等.反卫星武器研制状况的分析研究.系统工程与电子技术,1994,16(1)：38-51.
[6]刘绍球,高淑霞,唐仲藩等.空间对抗发展态势研究[J].系统工程与电子技术,1995,12.17(12)：17-27.
[7]张光义,王德纯.弹道导弹防御系统中的预警探测雷达[J].系统工程与电子技术,1996,18(5)：28-31.
[8]朱毅麟.空间碎片的观测、模型与缓减.863航天技术通讯,1998,6.1-25.
[9]戴征坚,郁文贤,胡卫东等.空间目标的雷达识别技术.系统工程与电子技术,2000,22(3)：19-22.
[10]Raup R. Bayesian Characterization and Detection of Rare Binary Signature Events. AD-A221116,1990.
[11]黄鸿勋译.观测外层空间目标用的跟踪和成像雷达系统.电子工程信息,1997(9)：1-11.
[12]黄小红,姜卫东.空间目标RCS序列周期性判定与提取[J].航天电子对抗,2005.21(2)：29-30.
[13]黄小红,邱兆坤,陈曾平.空间目标RCS序列的分形分析[J].中国空间科学技术.2005,2(1)：33-36.
[14]周文辉,胡卫东,计科峰等.基于雷达观测的空间目标识别仿真软件的设计与实现[J].系统仿真学报.2002,2,14(2)：238-242.
[15]刘先康,高梅国,傅雄军.基于HRRP偶数阶中心矩特征的卫星目标识别.电子器件.2007,10.30(5)：1626-1629.
[16]陈文彤,刘朝军,张汉华等.一种基于窄带雷达的低轨空间目标识别方法.中国空间科学技术.2006,8.(4)：48-54.
[17]马君国.空间雷达目标特征提取与识别方法研究.国防科学技术大学,博士学位论文.2006.
[18]刘宏伟,杜兰,袁莉等.雷达高分辨距离像目标识别研究进展.电子与信息学报.2005,8.27(8)：1328-1333.
[19]廖学军.基于高分辨距离像的雷达目标识别.[博士学位论文],西安：西安电子科技大学,1999.
[20]Chandran V, Elgar S L. Pattern recognition using invariants defined from higher order spectra one-dimensional inputs. IEEE Trans. on Signal Processing,1993,41(1). pp205-212.
[21]Tugnait J K. Detection of non-Gaussian signals using integrated polyspectrum. IEEE Trans. on Signal Processing.1994,42(11). pp3137-3149.
[22]Liao X, Bao Z. Circularly integrated bispectra:novel shift invariant feature for high-resolution radar target recognition. IEE Electronics Letters,1999,34(19). pp1879-1880.
[23]Zhang X, Shi Y, Bao Z. A new feature vector using selected bispectra for signal classification with application in radar target recognition. IEEE Trans. on Signal Processing.2001.49(9). pp1875-1885.
[24]李欣.基于RCS的空间目标识别技术.南京理工大学.硕士学位论文.2006.6..
[25]杜亚娟,潘泉,张洪才.雷达目标识别方法概述.火控雷达技术.1998.27(4)：1-4.
[26]边肇祺,张学工等.模式识别.第二版.北京：清华大学出版社,2000.
[27]袁莉.基于中心矩特征的雷达HRRP自动目标识别.电子学报.2004,12.32(12)：2078-2081.
[28]Peter N. Belhumeur, Joao P. Hespanha, David J. kriefina. Eigenfaces vs. Fisherfaces, Recognition Using Class Specific Linear Projection. IEEE Transactions On Pattern Analysis And Machine Intelligence, July 1997, vol.19, no.7, pp711-720.
[29]徐之海,冯华君,李奇等.基于Karhunen-Loeve变换的人脸识别研究.光学工程.2001,28(6)：48-51.
[30]阎敬文,沈贵明.基于KT/WT/WTVQ的三维多光谱数据压缩方法.厦门大学学报(自然科学版).2001,40(5)：1051-1055.
[31]李光,于盛林.基于小波变换的拟I/F信号生成.数据采集预处理.2003,18(1):119-122.
[32]Kim K T, Seo D K, Kim H T. Efficient radar target recognition using the MUSIC algorithm and invariant features[J].IEEE Trans A P,2002,50(3). pp325-337.
[33]吴秋荣.雷达高分辨距离像目标识别方法研究.成都：电子科技大学.硕士学位论文.2007,5.
[34]王雪娇.基于雷达距离像的目标识别研究.浙江大学.硕士论文.2007,5.
[35]Burges C J C. A Tutorial on Support Vector Machines for Pattern Recognition[J]. Data Mining and Knowledge Discovery,1998,2(2). ppl21-167.
[36]Vapnik V N. The Nature of Statistical Learning Theory [M]. New York:Springer-Verlag, 1995.
[37]Richard O. Duda, Peter E. Hart, Duda G. Stork.李宏东,姚天翔等译.模式分类.(中文版)第二版.北京：机械工业出版社,中信出版社,2003.
[38]胡利平,刘宏伟,吴顺君.一种新的SAR图像目标识别预处理方法.西安电子科技大学 学报(自然科学版).2007,10.34(5)：733-737.
[39]Fisher R.A. The Use of Multiple Measures in Taxonomic Problems. Ann. Eugenics.1936, Vol.7, pp179-188.
[40]王蕴红,刘国岁,李玺等,基于短时傅里叶变换及奇异值特征提取的目标识别方法,信号处理,第14卷,第2期,1998：ppl23-127
[41]温富喜,刘宏伟.基于中心矩特征的空间目标识别方法.雷达科学与技术.2007.2,5(1)：pp8-12.
[2]李海顺,董天临,曹毅强.空间测量雷达目标分类与识别算法及其能力分析.舰船电子工程.2005,3,25(3)：106-109.
[3]李玉书.空间目标监视雷达技术探讨.飞行器测控学报.2003.22(4)：62-66.
[4]魏晨曦,俄罗斯的空间目标监视、识别、探测与跟踪系统.中国航天.2006,8：39-41.
[5]刘绍球,高淑霞,高永浩等.反卫星武器研制状况的分析研究.系统工程与电子技术,1994,16(1)：38-51.
[6]刘绍球,高淑霞,唐仲藩等.空间对抗发展态势研究[J].系统工程与电子技术,1995,12.17(12)：17-27.
[7]张光义,王德纯.弹道导弹防御系统中的预警探测雷达[J].系统工程与电子技术,1996,18(5)：28-31.
[8]朱毅麟.空间碎片的观测、模型与缓减.863航天技术通讯,1998,6.1-25.
[9]戴征坚,郁文贤,胡卫东等.空间目标的雷达识别技术.系统工程与电子技术,2000,22(3)：19-22.
[10]Raup R. Bayesian Characterization and Detection of Rare Binary Signature Events. AD-A221116,1990.
[11]黄鸿勋译.观测外层空间目标用的跟踪和成像雷达系统.电子工程信息,1997(9)：1-11.
[12]黄小红,姜卫东.空间目标RCS序列周期性判定与提取[J].航天电子对抗,2005.21(2)：29-30.
[13]黄小红,邱兆坤,陈曾平.空间目标RCS序列的分形分析[J].中国空间科学技术.2005,2(1)：33-36.
[14]周文辉,胡卫东,计科峰等.基于雷达观测的空间目标识别仿真软件的设计与实现[J].系统仿真学报.2002,2,14(2)：238-242.
[15]刘先康,高梅国,傅雄军.基于HRRP偶数阶中心矩特征的卫星目标识别.电子器件.2007,10.30(5)：1626-1629.
[16]陈文彤,刘朝军,张汉华等.一种基于窄带雷达的低轨空间目标识别方法.中国空间科学技术.2006,8.(4)：48-54.
[17]马君国.空间雷达目标特征提取与识别方法研究.国防科学技术大学,博士学位论文.2006.
[18]刘宏伟,杜兰,袁莉等.雷达高分辨距离像目标识别研究进展.电子与信息学报.2005,8.27(8)：1328-1333.
[19]廖学军.基于高分辨距离像的雷达目标识别.[博士学位论文],西安：西安电子科技大学,1999.
[20]Chandran V, Elgar S L. Pattern recognition using invariants defined from higher order spectra one-dimensional inputs. IEEE Trans. on Signal Processing,1993,41(1). pp205-212.
[21]Tugnait J K. Detection of non-Gaussian signals using integrated polyspectrum. IEEE Trans. on Signal Processing.1994,42(11). pp3137-3149.
[22]Liao X, Bao Z. Circularly integrated bispectra:novel shift invariant feature for high-resolution radar target recognition. IEE Electronics Letters,1999,34(19). pp1879-1880.
[23]Zhang X, Shi Y, Bao Z. A new feature vector using selected bispectra for signal classification with application in radar target recognition. IEEE Trans. on Signal Processing.2001.49(9). pp1875-1885.
[24]李欣.基于RCS的空间目标识别技术.南京理工大学.硕士学位论文.2006.6..
[25]杜亚娟,潘泉,张洪才.雷达目标识别方法概述.火控雷达技术.1998.27(4)：1-4.
[26]边肇祺,张学工等.模式识别.第二版.北京：清华大学出版社,2000.
[27]袁莉.基于中心矩特征的雷达HRRP自动目标识别.电子学报.2004,12.32(12)：2078-2081.
[28]Peter N. Belhumeur, Joao P. Hespanha, David J. kriefina. Eigenfaces vs. Fisherfaces, Recognition Using Class Specific Linear Projection. IEEE Transactions On Pattern Analysis And Machine Intelligence, July 1997, vol.19, no.7, pp711-720.
[29]徐之海,冯华君,李奇等.基于Karhunen-Loeve变换的人脸识别研究.光学工程.2001,28(6)：48-51.
[30]阎敬文,沈贵明.基于KT/WT/WTVQ的三维多光谱数据压缩方法.厦门大学学报(自然科学版).2001,40(5)：1051-1055.
[31]李光,于盛林.基于小波变换的拟I/F信号生成.数据采集预处理.2003,18(1):119-122.
[32]Kim K T, Seo D K, Kim H T. Efficient radar target recognition using the MUSIC algorithm and invariant features[J].IEEE Trans A P,2002,50(3). pp325-337.
[33]吴秋荣.雷达高分辨距离像目标识别方法研究.成都：电子科技大学.硕士学位论文.2007,5.
[34]王雪娇.基于雷达距离像的目标识别研究.浙江大学.硕士论文.2007,5.
[35]Burges C J C. A Tutorial on Support Vector Machines for Pattern Recognition[J]. Data Mining and Knowledge Discovery,1998,2(2). ppl21-167.
[36]Vapnik V N. The Nature of Statistical Learning Theory [M]. New York:Springer-Verlag, 1995.
[37]Richard O. Duda, Peter E. Hart, Duda G. Stork.李宏东,姚天翔等译.模式分类.(中文版)第二版.北京：机械工业出版社,中信出版社,2003.
[38]胡利平,刘宏伟,吴顺君.一种新的SAR图像目标识别预处理方法.西安电子科技大学 学报(自然科学版).2007,10.34(5)：733-737.
[39]Fisher R.A. The Use of Multiple Measures in Taxonomic Problems. Ann. Eugenics.1936, Vol.7, pp179-188.
[40]王蕴红,刘国岁,李玺等,基于短时傅里叶变换及奇异值特征提取的目标识别方法,信号处理,第14卷,第2期,1998：ppl23-127
[41]温富喜,刘宏伟.基于中心矩特征的空间目标识别方法.雷达科学与技术.2007.2,5(1)：pp8-12.