«上一篇
 文章快速检索 高级检索

 雷达学报  2017, Vol. 6 Issue (5): 503-513  DOI: 10.12000/JR17047 0

引用本文

Dou Fangzheng, Diao Wenhui, Sun Xian, et al. Aircraft reconstruction in high resolution sar images using deep shape prior[J]. Journal of Radars, 2017, 6(5): 503-513. DOI: 10.12000/JR17047.

文章历史

(中国科学院电子学研究所   北京   100190)
(中国科学院大学   北京   100190)

Aircraft Reconstruction in High Resolution SAR Images Using Deep Shape Prior
Dou Fangzheng①②, Diao Wenhui, Sun Xian, Zhang Yue, Fu Kun
(Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China)
(University of Chinese Academy of Sciences, Beijing 100190, China)
Foundation Item: The National Natural Science Foundation of China (61331017)
Abstract: Object reconstruction is of vital importance in Synthetic Aperture Radar (SAR) image analysis. In this paper, we propose a novel method based on shape prior to reconstruct aircraft in high resolution SAR images. The method mainly contains two stages. In the shape prior modeling stage, a generative deep learning method is used to model deep shape priors; a novel framework is then proposed in the reconstruction stage, which integrates the shape priors in the process of reconstruction. Specifically, to address the issue of object rotation, a novel pose estimation method is proposed to obtain candidate poses, which avoids making an exhaustive search for each pose. In addition, an energy function combining a scattering region term and a shape prior term is proposed; this is optimized via an iterative optimization algorithm to achieve the goal of object reconstruction. To the best of our knowledge, this is the first attempt made to reconstruct objects with complex shapes in SAR images using deep shape priors. Experiments are conducted on the dataset acquired by TerraSAR-X and results demonstrate the accuracy and robustness of the proposed method.
Key words: Synthetic Aperture Radar (SAR)    Object reconstruction    Shape prior    Deep Boltzmann machine
1 引言

2 数据描述与分析

 图 1 TerraSAR-X数据集部分切片展示 Fig.1 Part of the TerraSAR-X data set

 图 2 8种特定角度下10种不同种类的飞机目标黑白二值模板(从上到下角度依次是0°, 315°, 270°, 225°, 180°, 135°, 90°和45°) Fig.2 Original black and white binary templates of ten types of aircrafts in 8 different poses (Poses, from top to bottom, are 0°, 315°, 270°, 225°, 180°, 135°, 90° and 45°)

3 形状先验模型学习

 ${x_{\rm{c}}} = \frac{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {{x_{i,j}}S({x_{i,j}})} } }}{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {S({x_{i,j}})} } }}, \ {y_{\rm{c}}} = \frac{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {{y_{i,j}}S({y_{i,j}})} } }}{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {S({y_{i,j}})} } }}$ (1)

 $\left. \begin{array}{l}{s_x} = {\left( {\frac{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {{{({x_{i,j}} - {x_{\rm{c}}})}^2}S({x_{i,j}})} } }}{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {S({x_{i,j}})} } }}} \right)^{\scriptsize\displaystyle\frac{1}{2}}}\\{s_y} = {\left( {\frac{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {{{({y_{i,j}} - {y_{\rm{c}}})}^2}S({y_{i,j}})} } }}{{\displaystyle\sum\limits_{i = 1}^m {\sum\limits_{j = 1}^n {S({y_{i,j}})} } }}} \right)^{\scriptsize\displaystyle\frac{1}{2}}}\end{array} \right\}$ (2)

 图 3 形状对齐结果(第1行和第2行分别表示0°姿态下原始形状模板和形状对齐后的结果图像) Fig.3 Results of shape alignment (The first and second rows represent the original templates and the results of shape alignment respectively)

 图 4 3层深度玻尔兹曼机网络结构 Fig.4 The structure of three-layered deep Boltzmann machine
4 基于形状先验的目标重建

 图 5 本文方法的流程图 Fig.5 The framework of the proposed aircraft reconstruction method

 $v(S,D) = \mathop {\max }\limits_S^4 \left\{ {\mathop {\max }\limits_{D = - N}^N {\Big |}{\rm corr}\left( {{X},{{X}_{S,D}}} \right){\Big |}} \right\}$ (3)

 $\begin{array}{c}E({q},{{h}^{{1}}},{{h}^{{2}}},φ ;{Θ} ) = \underbrace {\parallel \! \nabla {q}\!{\parallel _{\rm{e}}} + \alpha {{q}^{\rm T}}{s}}_{{\rm{scattering {\scriptsize{-}} term}}} - \underbrace {\beta ({{q}^{\rm T}}{W}_φ^1{{h}^1} + {{h}^1}^{^{\large\rm{T}}}{{W}^2}{{h}^2} + {{a}^1}^{^{\large\rm T}}{{h}^1} + {{a}^2}^{^{\large\rm T}}{{h}^2} + {{q}^{\rm T}}{b})}_{{\rm{shape {\scriptsize{-}} term}}}\end{array}$ (4)

 $φ = \left[\!\! {\begin{array}{*{20}{c}}1&0&x\\0&1&y\end{array}} \!\!\right]\left[\!\! {\begin{array}{*{20}{c}}h&0&0\\0&h&0\\0&0&1\end{array}} \!\!\right]\left[ \!\!{\begin{array}{*{20}{c}}{\cos \theta }&{ - \sin \theta }&0\\{\sin\theta }&\quad {\cos \theta }&0\\0&\quad 0&1\end{array}} \!\!\right]$ (5)

 $\parallel \nabla {q}{\parallel _{\rm{e}}} = \int_{Ω} {{{r}_{\rm{e}}} \cdot |\nabla {q}|} {\mathop{\rm d}\nolimits} {x}$ (6)

 ${s} = {({c_1} - {u})^2} - {({c_2} - {u})^2}$ (7)

5 函数优化与目标重建

6 实验结果与分析

 图 6 姿态估计结果 Fig.6 Results of the pose estimation

 ${\rm{PMP}} = \left(1 + \frac{{|{B_{\rm{o}}} \cap {B_{\rm{s}}}| + |{F_{\rm{o}}} \cap {F_{\rm{s}}}|}}{{|{B_{\rm{o}}}| + |{F_{\rm o}}|}}\right) \times 100\%$ (8)

 图 7 不同目标重建方法的结果对比(第1行与第2行分别对应提取目标区域的轮廓图和提取到的目标区域图) Fig.7 Comparison of reconstruction results of different methods (The first and the second rows correspond to the contour of the extracted aircraft area and the extracted aircraft area)

 图 8 目标重建结果 Fig.8 Reconstruction results

7 结束语

 [1] 邓云凯, 赵凤军, 王宇. 星载SAR技术的发展趋势及应用浅析[J]. 雷达学报, 2012, 1(1): 1-10. Deng Yunkai, Zhao Fengjun and Wang Yu. Brief analysis on the development and application of spaceborne SAR[J]. Journal of Radars, 2012, 1(1): 1-10. (0) [2] Chen Jiehong, Zhang Bo and Wang Chao. Backscattering feature analysis and recognition of civilian aircraft in TerraSAR-X images[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(4): 796-800. DOI:10.1109/LGRS.2014.2362845 (0) [3] Chang Y L, Chiang C Y and Chen K S. SAR image simulation with application to target recognition[J]. Progress in Electromagnetics Research, 2011, 119: 35-57. DOI:10.2528/PIER11061507 (0) [4] Tang Kan, Sun Xian, Sun Hao, et al.. A geometrical-based simulator for target recognition in high-resolution SAR images[J]. IEEE Geoscience and Remote Sensing Letters, 2012, 9(5): 958-962. DOI:10.1109/LGRS.2012.2187426 (0) [5] Zhang Yue, Sun Xian, Thiele A, et al.. Stochastic geometrical model and Monte Carlo optimization methods for building reconstruction from InSAR data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 108: 49-61. DOI:10.1016/j.isprsjprs.2015.06.004 (0) [6] 王思雨, 高鑫, 孙皓. 基于卷积神经网络的高分辨率SAR图像飞机目标检测方法[J]. 雷达学报, 2017, 6(2): 195-203. Wang Siyu, Gao Xin, Sun Hao, et al.. An aircraft detection method based on convolutional neural networks in high-resolution sar images[J]. Journal of Radars, 2017, 6(2): 195-203. (0) [7] 杜康宁, 邓云凯, 王宇. 基于多层神经网络的中分辨SAR图像时间序列建筑区域提取[J]. 雷达学报, 2016, 5(4): 410-418. Du Kangning, Deng Yunkai, Wang Yu, et al.. Medium Resolution SAR Image Time-series Built-up Area Extraction Based on Multilayer Neural Network[J]. Journal of Radars, 2016, 5(4): 410-418. (0) [8] 田壮壮, 占荣辉, 胡杰民. 基于卷积神经网络的SAR图像目标识别研究[J]. 雷达学报, 2016, 5(3): 320-325. Tian Zhuangzhuang, Zhan Ronghui, Hu Jiemin, et al.. SAR ATR Based on Convolutional Neural Network[J]. Journal of Radars, 2016, 5(3): 320-325. (0) [9] 徐丰, 王海鹏, 金亚秋. 深度学习在SAR目标识别与地物分类中的应用[J]. 雷达学报, 2017, 6(2): 136-148. Xu Feng, Wang Haipeng and Jin Yaqiu. Deep learning as applied in sar target recognition and terrain classification[J]. Journal of Radars, 2017, 6(2): 136-148. (0) [10] Salakhutdinov R and Hinton G. Deep Boltzmann machines[J]. Journal of Machine Learning Research, 2009, 5(2): 448-455. (0) [11] Zhang Zhengdong, Liang Xiao, Ganesh A, et al.. TILT: Transform invariant low-rank textures[C]. Proceedings of Asian Conference on Computer Vision, Berlin, Heidelberg, 2010: 314–328. (0) [12] Goldstein T, Bresson X and Osher S. Geometric applications of the split bregman method: Segmentation and surface reconstruction[J]. Journal of Scientific Computing, 2010, 45(1): 272-293. (0) [13] Eslami S M A, Heess N, Williams C K I, et al.. The shape boltzmann machine: A strong model of object shape[J]. International Journal of Computer Vision, 2014, 107(2): 155-176. DOI:10.1007/s11263-013-0669-1 (0) [14] Liu Ge, Sun Xian, Fu Kun, et al.. Interactive geospatial object extraction in high resolution remote sensing images using shape-based global minimization active contour model[J]. Pattern Recognition Letters, 2013, 34(10): 1186-1195. DOI:10.1016/j.patrec.2013.03.031 (0) [15] Chen Fei, Yu Huimin, Hu R, et al.. Deep learning shape priors for object segmentation[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Portland, OR, USA, 2013: 1870–1877. (0) [16] Kuttikkad S and Chellappa R. Non-gaussian CFAR techniques for target detection in high resolution SAR images[C]. Proceedings of IEEE International Conference on Image Processing, Austin, TX, USA, 1994: 910–914. (0) [17] Cremers D, Schmidt F R, and Barthel F. Shape priors in variational image segmentation: Convexity, Lipschitz continuity and globally optimal solutions[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Anchorage, AK, USA, 2008: 1–6. (0) [18] Jones III G and Bhanu B. Recognizing articulated objects in SAR images[J]. Pattern Recognition, 2001, 34(2): 469-485. DOI:10.1016/S0031-3203(99)00218-6 (0) [19] Boykov Y Y and Jolly M P. Interactive graph cuts for optimal boundary & region segmentation of objects in N-D images[C]. Proceedings of IEEE International Conference on Computer Vision, Vancouver, BC, Canada, 2001: 105–112. (0) [20] Wu Qichang, Diao Wenhui, Dou Fangzheng, et al.. Shape-based object extraction in high-resolution remote-sensing images using deep Boltzmann machine[J]. International Journal of Remote Sensing, 2016, 37(24): 6012-6022. DOI:10.1080/01431161.2016.1253897 (0)