属性散射中心匹配及其在SAR目标识别中的应用

丁柏圆; 文贡坚; 余连生; 马聪慧

doi:10.12000/JR16104

属性散射中心匹配及其在SAR目标识别中的应用

DOI: 10.12000/JR16104

丁柏圆^①, ,,
文贡坚^①,
余连生^②,
马聪慧^①

①.
国防科学技术大学ATR重点实验室长沙 410073
②.
中国天绘卫星中心北京 102102

基金项目:

教育部新世纪优秀人才支持计划 NCET-11-0866

详细信息

作者简介:
丁柏圆 (1990–), 男, 博士研究生, 研究方向为SAR自动目标识别。E-mail:Dingbaiyuan_NUDT@163.com

文贡坚 (1972–), 男, 教授, 博士生导师, 研究方向为遥感图像处理

余连生 (1984–), 男, 助理工程师, 研究方向为摄影测量与遥感

马聪慧 (1987–), 女, 1987年生, 博士研究生, 研究方向为SAR自动目标识别

通讯作者:
丁柏圆, E-mail:Dingbaiyuan_NUDT@163.com

中图分类号: TN957
计量
- 文章访问数:
- HTML全文浏览量:
- PDF下载量:
- 被引次数: 0
出版历程
- 收稿日期: 2016-09-15
- 修回日期: 2016-11-25
- 网络出版日期: 2017-04-28

Matching of Attributed Scattering Center and Its Application to Synthetic Aperture Radar Automatic Target Recognition

Ding Baiyuan^{①
, ,},
Wen Gongjian^①,
Yu Liansheng^②,
Ma Conghui^①

①.
ATR Key Laboratory National University of Defense Technology, Changsha 410073, China
②.
TH Center of China, Beijing 102102, China

Funds:

The Program for New Century Excellent Talents in University NCET-11-0866

摘要

摘要: 属性散射中心是合成孔径雷达 (Synthetic Aperture Radar, SAR) 图像的一个重要特征。该文提出了一种属性散射中心匹配方法并将其运用于SAR目标识别中。该方法首先基于属性散射中心模型提取待识别SAR图像和模板SAR图像的属性散射中心，进而采用Hungarian算法实现散射中心的匹配。在建立的匹配关系的基础上，设计了一种稳健的散射中心匹配度度量方法计算待识别散射中心与各类模板散射中心的匹配度。该匹配度准则充分考虑了单个散射中心强弱、匹配对强弱以及漏警、虚警带来的影响，对于散射中心集的匹配度的评价更为全面。基于Moving and Stationary Target Acquisition and Recognition (MSTAR) 数据集的实验验证了方法的有效性。
- 合成孔径雷达 /
- 目标识别 /
- 属性散射中心匹配 /
- Hungarian算法 /
- 匹配度度量方法
Abstract: Attributed scattering center is one of important features of Synthetic Aperture Radar (SAR) image. In this paper, a method for the matching of attributed scattering centers and its application to SAR target recognition is proposed. First, the attributed scattering centers of the test SAR image and template SAR images are extracted on the basis of the attributed scattering model. Second, the Hungarian algorithm is employed to match the two scattering center sets. Based on the one to one correspondence, we design a new similarity measure to evaluate the similarity between the two scattering center sets that will decide the target type of the test image. The similarity measure considers the effects of each individual scattering center, single matching pair, and missing alarms and false alarms; thus, it is more comprehensive. The experiment based on moving and stationary target acquisition and recognition database demonstrates the validity of the proposed method.
- Synthetic Aperture Radar (SAR) /
- Target recognition /
- Matching of attributed scattering center /
- Hungarian algorithm /
- Similarity measure

HTML全文

图 1 一幅BMP2 SAR图像的参数估计重构结果 (动态范围40 dB)

Figure 1. The reconstruction of a BMP2 SAR image (Dynamic range: 40 dB)

下载: 全尺寸图片幻灯片

图 2 Hungarian算法实现散射中心匹配

Figure 2. The matching of attributed scattering centers by Hungarian algorithm

下载: 全尺寸图片幻灯片

图 3 门限计算

Figure 3. The calculation of the threshold

下载: 全尺寸图片幻灯片

图 4 “强匹配对”的权值计算

Figure 4. The weight of the "strong assignments"

下载: 全尺寸图片幻灯片

图 5 漏警和虚警的权值计算

Figure 5. The weight of MAs and FAs

下载: 全尺寸图片幻灯片

图 6 本文方法的基本流程

Figure 6. The general procedure of the proposed method

下载: 全尺寸图片幻灯片

图 7 本文方法在不同模型阶数的识别性能

Figure 7. The recognition performance under different model orders

下载: 全尺寸图片幻灯片

图 8 不同分辨率下的MSTAR图像

Figure 8. The MSTAR images under different resolutions

下载: 全尺寸图片幻灯片

图 9 不同分辨率下的识别率

Figure 9. The recognition performance under different resolutions

下载: 全尺寸图片幻灯片

图 10 不同信噪比下的MSTAR图像

Figure 10. The reconstructed MSTAR images under different SNRs

下载: 全尺寸图片幻灯片

图 11 不同信噪比下的识别性能

Figure 11. The recognition performance under different SNRs

下载: 全尺寸图片幻灯片

表 1 分配代价矩阵

Table 1. The cost matrix for Hungarian matching

	Y₁	Y₂	$\begin{array}{*{20}{c}}\cdots \end{array}$	Y_n	MA
X₁	C₁₁	C₁₂	$\begin{array}{*{20}{c}}\cdots \end{array}$	C_1n	M₁	$\infty$	$\begin{array}{*{20}{c}}\cdots \end{array}$	$\infty$
X₂	C₂₁	C₂₂	$\begin{array}{*{20}{c}}\cdots \end{array}$	C_2n	$\infty$	M₂	$\begin{array}{*{20}{c}}\cdots \end{array}$	$\infty$
$\vdots$	$\vdots$	$\vdots$	$\ddots$	$\vdots$	$\vdots$	$\vdots$	$\ddots$	$\vdots$
X_m	C_m1	C_m2	$\begin{array}{*{20}{c}}\cdots \end{array}$	C_mn	$\infty$	$\infty$	$\begin{array}{*{20}{c}}\cdots \end{array}$	M_m
FA	F₁	$\infty$	$\ldots$	$\infty$	$\infty$	$\infty$	$\ldots$	$\infty$
	$\infty$	F₂	$\ldots$	$\infty$	$\infty$	$\infty$	$\ldots$	$\infty$
	$\vdots$	$\vdots$	$\ddots$	$\vdots$	$\vdots$	$\vdots$	$\ddots$	$\vdots$
	$\infty$	$\infty$	$\ldots$	F_n	$\infty$	$\infty$	$\ldots$	$\infty$

下载: 导出CSV

表 2 属性的不确定性建模

Table 2. The modeling of attribute uncertainty

属性类别	均值	方差
方位向x	x₀	${\sigma ^2} = {\rm{CR}}{{\rm{R}}^2}$
距离向y	y₀	${\sigma ^2} = {\rm{R}}{{\rm{R}}^2}$
归一化幅度\|A\|	\|A₀\|	${\sigma ^2} = 0.1$

下载: 导出CSV

表 3 模板集和测试集

Table 3. The template set and testing set

模板集	样本数量	测试集	样本数量
BMP2(SN_C21)	233	BMP2(SN_C21)	195
BMP2(SN_9566)	232	BMP2(SN_9566)	196
BMP2(SN_9563)	233	BMP2(SN_9563)	196
BTR70(SN_C71)	233	BTR70(SN_C71)	196
T72(SN_132)	232	T72(SN_132)	196
T72(SN_812)	231	T72(SN_812)	195
T72(SN_S7)	228	T72(SN_S7)	191

下载: 导出CSV

表 4 本文方法的识别结果

Table 4. The recognition result of the proposed method

目标类型	识别结果			P_c(%)
目标类型	BMP2	BTR70	T72	P_c(%)
BMP2(SN_9563)	191	2	2	97.95
BMP2(SN_9566)	195	1	0	99.49
BMP2(SN_C21)	192	2	2	97.96
BTR70	5	188	3	95.92
T72(SN_132)	8	1	187	96.91
T72(SN_812)	1	0	194	99.49
T72(SN_S7)	4	1	186	97.31
Average (%)	97.88

下载: 导出CSV

表 5 3种方法的识别性能对比

Table 5. The comparison of the three methods

方法	识别率 (%)	消耗时间 (ms)
本文方法	97.88	10.2
方法一	93.46	3.5
方法二	96.92	20.3

下载: 导出CSV

参考文献(16)

[1]	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: 22–24. Huang Pei-kang, Yin Hong-cheng, and Xu Xiao-jian. Radar Target Signature[M]. Beijing: Publishing House of Electronics Industry, 2005: 22–24.
[2]	Keller J B. Geometrical theory of diffraction[J]. Journal of the Optical Society of America, 1962: 52(2): 116–130. doi: 10.1364/JOSA.52.000116
[3]	Gerry M J, Potter L C, and Gupta I J. A parametric model for synthetic aperture radar measurements[J]. IEEE Transactions on Antennas and Propagation, 1999, 47(7): 1179–1188. doi: 10.1109/8.785750
[4]	Chiang H-C, Moses R L, and Potter L C. Model-based classification of radar images[J]. IEEE Transactions on Information Theory, 2000, 46(5): 1842–1854. doi: 10.1109/18.857795
[5]	唐涛, 粟毅.散射中心特征序贯匹配的SAR图像目标识别方法[J].系统工程与电子技术, 2012, 34(6): 1131–1135. http://www.cnki.com.cn/Article/CJFDTOTAL-XTYD201206011.htm Tang Tao and Su Yi. Object recognition in SAR imagery using sequential feature matching of scattering centers[J]. System Engineering and Electronics, 2012, 34(6): 1131–1135. http://www.cnki.com.cn/Article/CJFDTOTAL-XTYD201206011.htm
[6]	Sirui Tian, Kuiying Yin, Chao Wang, et al.. An SAR ATR method based on scattering center feature and bipartite graph matching[J]. IETE Technical Review, 2015, 32(5): 364–375. doi: 10.1080/02564602.2015.1019941
[7]	Bhanu B and Lin Y Q. Stochastic models for recognition of occluded targets[J]. Pattern Recognition, 2003, 36(12): 2855–2873. doi: 10.1016/S0031-3203(03)00182-1
[8]	Dungan K E and Potter L C. Classifying transformationvariant attributed patterns[J]. Pattern Recognition, 2010, 43(11): 3805–3816. doi: 10.1016/j.patcog.2010.05.033
[9]	Dungan K E and Potter L C. Classifying sets of attributed scattering centers using a hash coded database[C]. Proceedings of Algorithms for Synthetic Aperture Radar Imagery XVII, SPIE, Florida, 2010: 7737R01–7737R11.
[10]	Zhou Jianxiong, Shi Zhiguang, Chen Xiao, et al.. Automatic target recognition of SAR images based on global scattering center model[J]. IEEE Transactions on Geosciences and Remote Sensing, 2011, 49(10): 3713–3729. doi: 10.1109/TGRS.2011.2162526
[11]	Kim Taejoon and Dong Miaomiao. An iterative Hungarian method to joint relay selection and resource allocation for D2D communications[J]. IEEE Wireless Communications Letters, 2014, 3(6): 625–629. doi: 10.1109/LWC.2014.2338318
[12]	Li D, Zhang G, Wu Z, et al.. An edge embedded markerbased watershed algorithm for high spatial resolution remote sensing image segmentatio[J]. IEEE Transactions on Image Processing, 2010, 19(10): 2781–2787. doi: 10.1109/TIP.2010.2049528
[13]	Jing M, Zhou X, and Qi C. Quasi-Newton Iterative Projection algorithm for sparse recovery[J]. Neurocomputing, 2014, 144: 169–173. doi: 10.1016/j.neucom.2014.04.055
[14]	Chen J, Li Y, Wang J, et al.. Adaptive CLEAN algorithm for millimetre wave synthetic aperture imaging[J]. IET Image Processing, 2015, 9(3): 218–225. doi: 10.1049/iet-ipr.2014.0443
[15]	陶勇, 胡卫东.基于图像域的属性散射中心分析[J].信号处理, 2009, 25(10): 1510–1514. doi: 10.3969/j.issn.1003-0530.2009.10.004 Tao Yong and Hu Wei-Dong. Analysis of attributed scattering center based on image domain[J]. Signal Processing, 2009, 25(10): 1510–1514. doi: 10.3969/j.issn.1003-0530.2009.10.004
[16]	徐牧, 王雪松, 肖顺平.基于Hough变换与目标主轴提取的SAR图像目标方位角估计方法[J].电子与信息学报, 2007, 29(2): 370–374. http://www.cnki.com.cn/Article/CJFDTOTAL-DZYX200702024.htm Xu Mu, Wang Xue-song, and Xiao Shun-ping. Target aspect estimation in SAR imagery based on Hough transform and major axis extraction[J]. Journal of Electronic & Information Technology, 2007, 29(2): 370–374. http://www.cnki.com.cn/Article/CJFDTOTAL-DZYX200702024.htm