Polarimetric SAR ship detection based on superpixel and sparse reconstruction saliency
-
摘要: 极化SAR舰船检测是极化SAR系统的重要应用之一，现有的极化SAR舰船检测方法在具有强背景杂波的条件下容易将强杂波误检为目标，造成虚警；在多尺度舰船检测情况下小尺寸的舰船容易淹没在背景杂波中，造成小尺寸目标的漏检。针对上述问题，本文提出一种基于超像素与稀疏重构显著性的极化SAR舰船检测方法。该方法首先用超像素分割方法将大幅的极化SAR场景图像分割，在超像素级别上使用稀疏重构显著性方法保留含有舰船目标的超像素，再在这些保留下来的超像素中逐像素使用稀疏重构显著性检测方法，得到最终的舰船检测结果。本文选取强杂波场景和多尺度舰船检测场景的两个场景的ALOS-2卫星极化SAR数据进行对比实验，实验结果表明，本文方法在强杂波场景下品质因数达到94.87%，在多尺度舰船检测场景下品质因数达到94.05%。Abstract: Polarimetric SAR ship detection is an important application of the polarimetric SAR system. Existing polarimetric SAR ship detection methods are plagued by erroneous detection of strong clutter and missed detection of small targets in multiscale situations. Particularly, the existing methods easily detect strong clutter as the target under strong background clutter, resulting in false alarms; in the case of multiscale ship detection, small ships are easily submerged in background clutter, resulting in missed detection of small targets. To solve these problems, this paper proposes a polarimetric SAR ship detection method based on superpixels and sparse reconstruction saliency. This method has two stages. In the first stage, the large polarimetric SAR ship detection scene image is segmented using the superpixel segmentation method to obtain a superpixel image. With the superpixel as the basic unit, a saliency detection method based on sparse reconstruction is used to obtain the saliency value of each superpixel in the image. Then, the superpixels that may contain ship targets are retained using the sea surface ship density defined in this paper. Accordingly, in the first stage, the superpixel regions that may contain ship targets are obtained through superpixel segmentation and sparse reconstruction saliency detection. Next, in the second stage, a saliency detection method based on sparse reconstruction is used to obtain the saliency value of each pixel in these reserved superpixel regions. Finally, the global threshold segmentation method is used for the pixels in these regions to obtain the final detection results of ship targets. In this paper, two polarimetric SAR images of the ALOS-2 satellite with different scenes were selected for an experiment. One image contains strong clutter on the sea surface; the other contains ships of different sizes and many small ships. The experimental results show that the proposed method can well determine the superpixel regions that may contain ship targets in the first stage and successfully obtain the ship detection results in the second stage. In addition, in both scenarios, the classic constant false alarm rate (CFAR) methods and a saliency detection method are used for comparison with the proposed method. The experimental results show that the proposed method produces almost no false alarms because it is insensitive to strong clutter; moreover, this method rarely misses small ship targets in the multiscale ship detection scene. The figure of merit of the proposed method reaches 94.87% in the strong clutter scene and 94.05% in the multiscale ship detection scene.
-
Key words:
- polarimetric SAR /
- ship detection /
- sparse reconstruction /
- superpixel segmentation /
- saliency detection
-
图 1 基于稀疏重构的显著性目标检测流程图
Figure 1. Flowchart of object detection based on sparse reconstruction saliency
图 2 基于超像素与稀疏重构显著性的极化SAR舰船检测流程图
Figure 2. Flowchart of polarimetric SAR ship detection based on superpixel and sparse reconstruction saliency
图 3 SLIC超像素分割效果图. (a) SLIC; (b) 基于修正Wishart分布的SLIC
Figure 3. SLIC superpixel segmentation results: (a) SLIC; (b) SLIC based on the revised Wishart distribution
图 6 超像素检测结果图. (a) P=0.3; (b) P=0.1
Figure 6. Superpixel detection result: (a) P = 0.3; (b) P = 0.1
图 8 场景一和场景二的Pauli为彩色图与真值图. (a)场景一的Pauli伪彩色图; (b)场景一的真值图; (c)场景二的Pauli伪彩色图; (d)场景二的真值图
Figure 8. Pauli pseudo-color maps and truth maps of scene 1 and scene 2: (a) Pauli pseudo-color map of scene 1; (b) truth map for scene 1; (c) Pauli pseudo-color map of scene 2; (d) truth map for scene 2
图 9 场景一的舰船检测结果. (a) CA-CFAR; (b) OS-CFAR; (c) 显著性方法; (d) 本文方法; (e) 场景一的真值图
Figure 9. Ship detection results in scene 1: (a) CA-CFAR; (b) OS-CFAR; (c) saliency method; (d) our method; (e) truth map for scene 1
图 10 场景二的舰船检测结果. (a) CA-CFAR; (b) OS-CFAR; (c) 显著性方法; (d) 本文方法; (e) 场景二的真值图
Figure 10. Ship detection results in scene 2: (a) CA-CFAR; (b) OS-CFAR; (c) saliency method; (d) our method; (e) truth map for scene 2
表 1 场景一极化SAR舰船定量检测结果
Table 1. Polarization SAR ship quantitative detection results of scene 1
Method Nc NFA NM FoM/% CA-CFAR 37 3 2 88.10 OS-CFAR 36 5 3 81.82 Saliency method 35 0 4 89.74 Our method 37 0 2 94.87 
下载: 导出CSV
表 2 场景二极化SAR舰船定量检测结果
Table 2. Polarization SAR ship quantitative detection results of scene 2
Method Nc NFA NM FoM/% CA-CFAR 75 0 7 91.46 OS-CFAR 77 6 5 87.50 Saliency method 65 0 17 79.27 Our method 79 2 3 94.05
下载: 导出CSV
-
参考文献
[1] Wang X S, Chen S W. Polarimetric synthetic aperture radar interpretation and recognition: Advances and perspectives. J Radars, 2020, 9(2): 259 doi: 10.12000/JR19109王雪松, 陈思伟. 合成孔径雷达极化成像解译识别技术的进展与展望. 雷达学报, 2020, 9(2):259 doi: 10.12000/JR19109 [2] Du L, Wang Z C, Wang Y, et al. Survey of research progress on target detection and discrimination of single-channel SAR images for complex scenes. J Radars, 2020, 9(1): 34 doi: 10.12000/JR19104杜兰, 王兆成, 王燕, 等. 复杂场景下单通道SAR目标检测及鉴别研究进展综述. 雷达学报, 2020, 9(1):34 doi: 10.12000/JR19104 [3] Deng Y K, Yu W D, Zhang H, et al. Forthcoming spaceborne SAR development. J Radars, 2020, 9(1): 1 doi: 10.12000/JR20008邓云凯, 禹卫东, 张衡, 等. 未来星载SAR技术发展趋势. 雷达学报, 2020, 9(1):1 doi: 10.12000/JR20008 [4] Zhang J, Zhang X, Fan C Q, et al. Discussion on application of polarimetric synthetic aperture radar in marine surveillance. J Radars, 2016, 5(6): 596张杰, 张晰, 范陈清, 等. 极化SAR在海洋探测中的应用与探讨. 雷达学报, 2016, 5(6):596 [5] Wang X S. Status and prospects of radar polarimetry techniques. J Radars, 2016, 5(2): 119 doi: 10.12000/JR16039王雪松. 雷达极化技术研究现状与展望. 雷达学报, 2016, 5(2):119 doi: 10.12000/JR16039 [6] Liu T, Yang Z Y, Jiang Y N, et al. Review of ship detection in polarimetric synthetic aperture imagery. J Radars, 2021, 10(1): 1 doi: 10.12000/JR20155刘涛, 杨子渊, 蒋燕妮, 等. 极化SAR图像舰船目标检测研究综述. 雷达学报, 2021, 10(1):1 doi: 10.12000/JR20155 [7] Cui X C, Su Y, Chen S W. Polarimetric SAR ship detection based on polarimetric rotation domain features and superpixel technique. J Radars, 2021, 10(1): 35 doi: 10.12000/JR20147崔兴超, 粟毅, 陈思伟. 融合极化旋转域特征和超像素技术的极化SAR舰船检测. 雷达学报, 2021, 10(1):35 doi: 10.12000/JR20147 [8] Liu T, Yang Z Y, Yang J, et al. CFAR ship detection methods using compact polarimetric SAR in a K-Wishart distribution. IEEE J STARS, 2019, 12(10): 3737 [9] Pappas O, Achim A, Bull D. Superpixel-level CFAR detectors for ship detection in SAR imagery. IEEE Geosci Remote Sens Lett, 2018, 15(9): 1397 doi: 10.1109/LGRS.2018.2838263 [10] Gao G. A parzen-window-kernel-based CFAR algorithm for ship detection in SAR images. IEEE Geosci Remote Sens Lett, 2011, 8(3): 557 doi: 10.1109/LGRS.2010.2090492 [11] Huang Y L, Sun L, Guo L, et al. Ship detection algorithm based on spatially variant apodization sidelobe suppression and order statistic-constant false alarm rate. J Radars, 2020, 9(2): 335 doi: 10.12000/JR19082黄寅礼, 孙路, 郭亮, 等. 基于空间变迹滤波旁瓣抑制与有序统计恒虚警率的舰船检测算法. 雷达学报, 2020, 9(2):335 doi: 10.12000/JR19082 [12] Chen X, Sun J, Yin K Y, et al. An algorithm of ship target detection in SAR images based on cascaded CFAR. Mod Radar, 2012, 34(9): 50 doi: 10.3969/j.issn.1004-7859.2012.09.011陈祥, 孙俊, 尹奎英, 等. 基于CFAR级联的SAR图像舰船目标检测算法. 现代雷达, 2012, 34(9):50 doi: 10.3969/j.issn.1004-7859.2012.09.011 [13] An W T, Xie C H, Yuan X Z. An improved iterative censoring scheme for CFAR ship detection with SAR imagery. IEEE Trans Geosci Remote Sens, 2014, 52(8): 4585 doi: 10.1109/TGRS.2013.2282820 [14] Weiss M. Analysis of some modified cell-averaging CFAR processors in multiple-target situations. IEEE Trans Aerosp Electron Syst, 1982, AES-18(1): 102 doi: 10.1109/TAES.1982.309210 [15] Novak L M, Burl M C, Irving W W. Optimal polarimetric processing for enhanced target detection. IEEE Trans Aerosp Electron Syst, 1993, 29(1): 234 doi: 10.1109/7.249129 [16] Goldstein G B. False-alarm regulation in log-normal and weibull clutter. IEEE Trans Aerosp Electron Syst, 1973, AES-9(1): 84 doi: 10.1109/TAES.1973.309705 [17] Ritcey J A, Du H. Order statistic CFAR detectors for speckled area targets in SAR // Conference Record of the Twenty-Fifth Asilomar Conference on Signals, Systems & Computers. Pacific Grove, 1991: 1082 [18] Anastassopoulos V, Lampropoulos G A. Optimal CFAR detection in weibull clutter. IEEE Trans Aerosp Electron Syst, 1995, 31(1): 52 doi: 10.1109/7.366292 [19] Leng X G, Ji K F, Xing X W, et al. Area ratio invariant feature group for ship detection in SAR imagery. IEEE J Sel Top Appl Earth Obs Remote Sens, 2018, 11(7): 2376 doi: 10.1109/JSTARS.2018.2820078 [20] Fan W W, Zhou F, Bai X R, et al. Ship detection using deep convolutional neural networks for PolSAR images. Remote Sens, 2019, 11(23): 2862 doi: 10.3390/rs11232862 [21] Chang Y L, Anagaw A, Chang L N, et al. Ship detection based on YOLOv2 for SAR imagery. Remote Sens, 2019, 11(7): 786 doi: 10.3390/rs11070786 [22] Zhang X L, Zhang T W, Shi J, et al. High-speed and high-accurate SAR ship detection based on a depthwise separable convolution neural network. J Radars, 2019, 8(6): 841 doi: 10.12000/JR19111张晓玲, 张天文, 师君, 等. 基于深度分离卷积神经网络的高速高精度SAR舰船检测. 雷达学报, 2019, 8(6):841 doi: 10.12000/JR19111 [23] Tian Z Z, Zhan R H, Hu J M, et al. SAR ATR based on convolutional neural network. J Radars, 2016, 5(3): 320 doi: 10.12000/JR16037田壮壮, 占荣辉, 胡杰民, 等. 基于卷积神经网络的SAR图像目标识别研究. 雷达学报, 2016, 5(3):320 doi: 10.12000/JR16037 [24] Chen S W, Tao C S, Wang X S, et al. Polarimetric SAR targets detection and classification with deep convolutional neural network // 2018 Progress in Electromagnetics Research Symposium (PIERS-Toyama). Toyama, 2019: 2227 [25] Zhou F, Fan W W, Sheng Q Q, et al. Ship detection based on deep convolutional neural networks for polsar images // IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium. Valencia, 2018: 681 [26] Jin K, Chen Y L, Xu B, et al. A patch-to-pixel convolutional neural network for small ship detection with PolSAR images. IEEE Trans Geosci Remote Sens, 2020, 58(9): 6623 doi: 10.1109/TGRS.2020.2978268 [27] Zhai L, Li Y, Su Y. Inshore ship detection via saliency and context information in high-resolution SAR images. IEEE Geosci Remote Sens Lett, 2016, 13(12): 1870 doi: 10.1109/LGRS.2016.2616187 [28] Cui X C, Su Y, Chen S W. A saliency detector for polarimetric SAR ship detection using similarity test. IEEE J Sel Top Appl Earth Obs Remote Sens, 2019, 12(9): 3423 doi: 10.1109/JSTARS.2019.2925833 [29] Wright J, Yang A Y, Ganesh A, et al. Robust face recognition via sparse representation. IEEE Trans Pattern Anal Mach Intell, 2009, 31(2): 210 doi: 10.1109/TPAMI.2008.79 [30] Li X H, Lu H C, Zhang L H, et al. Saliency detection via dense and sparse reconstruction // 2013 IEEE International Conference on Computer Vision. Sydney, 2013: 2976 [31] Lu H C, Li X H, Zhang L H, et al. Dense and sparse reconstruction error based saliency descriptor. IEEE Trans Image Process, 2016, 25(4): 1592 doi: 10.1109/TIP.2016.2524198 [32] Borji A, Itti L. Exploiting local and global patch rarities for saliency detection // 2012 IEEE Conference on Computer Vision and Pattern Recognition. Providence, 2012: 478 [33] Achanta R, Shaji A, Smith K, et al. SLIC superpixels compared to state-of-the-art superpixel methods. IEEE Trans Pattern Anal Mach Intell, 2012, 34(11): 2274 doi: 10.1109/TPAMI.2012.120 [34] Qin F C, Guo J M, Lang F K. Superpixel segmentation for polarimetric SAR imagery using local iterative clustering. IEEE Geosci Remote Sens Lett, 2015, 12(1): 13 doi: 10.1109/LGRS.2014.2322960 [35] Li T, Liu Z, Xie R, et al. An improved superpixel-level CFAR detection method for ship targets in high-resolution SAR images. IEEE J Sel Top Appl Earth Obs Remote Sens, 2018, 11(1): 184 doi: 10.1109/JSTARS.2017.2764506 [36] Cui X C, Tao C S, Su Y, et al. PolSAR ship detection based on polarimetric correlation pattern. IEEE Geosci Remote Sens Lett, 2021, 18(3): 471 doi: 10.1109/LGRS.2020.2976477 -
点击查看大图
计量
- 文章访问数: 122
- HTML全文浏览量: 12
- PDF下载量: 9
- 被引次数: 0
