基于超密集特征金字塔网络的SAR图像舰船检测

doi:10.3969/j.issn.1001-506X.2020.10.09

摘要/Abstract

摘要：

针对星载合成孔径雷达(synthetic aperture radar, SAR)图像舰船目标检测困难的问题,提出了一种基于超密集特征金字塔网络的检测算法。首先,利用残差神经网络提取原始图像特征,构建特征图。其次,跨尺度连接多个特征层获取超密集特征金字塔,建立多尺度的高层语义特征映射,增强特征传播和重用。然后，再利用区域建议网络提取每层金字塔的候选区域输入检测网络。最后,通过融合候选区域及其周边上下文信息,将检测网络注意力集中至海域以抑制虚警,并为分类器计算置信度和边框回归提供补充信息。多组仿真实验证明,所提网络框架设定合理且检测性能优越。

关键词: 合成孔径雷达, 卷积神经网络, 超密集特征金字塔网络, 上下文信息

Abstract:

Aiming at the difficulty of ship target detection in space-borne synthetic aperture radar (SAR) images, a detection algorithm based on super dense feature pyramid networks is proposed. Firstly, residual neural network is used to extract features from original images and construct feature maps. Secondly, in order to enhance feature propagation and reuse, cross-scale feature layers are connected to obtain super dense feature pyramid and establish multi-scale high-level semantic feature mapping. Thirdly, candidate region is extracted from each layer of pdyramids by the regional proposal networks, and input into the detectim network. Finally, by fusing the candidate region and its surrounding contextual information to make the detector focus on the sea areas to suppress the false alarms, and provides supplementary information for the classifier to calculate confidence and bounding box regression. Simulation experiments show that the proposed network framework is reasonable and the detection performance is superior.

Key words: synthetic aperture radar (SAR), convolutional neural network (CNN), super dense feature pyramid networks, contextual information

中图分类号:

TP391

韩子硕, 王春平, 付强, 徐艳. 基于超密集特征金字塔网络的SAR图像舰船检测[J]. 系统工程与电子技术, 2020, 42(10): 2214-2222.

Zishuo HAN, Chunping WANG, Qiang FU, Yan XU. Ship detection in SAR images based on super dense feature pyramid networks[J]. Systems Engineering and Electronics, 2020, 42(10): 2214-2222.

图/表 14

图1

图2

图3

图4

图5

图6

图7

表1

表2

表3

表4

图8

表5

图9

参考文献 32

1	SHAO J Q, QU C W, LI J W. A performance analysis of convolutional neural network models in SAR target recognition[C]//Proc.of the SAR in Big Data Era: Models, Methods and Applications, 2017.
2	AO W , XU F , LI Y C , et al. Detection and discrimination of ship targets in complex background from spaceborne ALOS-2 SAR images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11 (2): 536- 550. doi: 10.1109/JSTARS.2017.2787573
3	WANG C L , BI F K , ZHANG W P , et al. An intensity-space domain CFAR method for ship detection in HR SAR images[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14 (4): 529- 533. doi: 10.1109/LGRS.2017.2654450
4	LENG X G, JI K F, FAN Q J, et al. A novel adaptive ship detection method for spaceborne SAR imagery[C]//Proc.of the IEEE International Geoscience and Remote Sensing Symposium, 2016.
5	ZHAO Z , JI K F , XING X W , et al. Ship surveillance by integration of space-borne SAR and AIS-review of current research[J]. Journal of Navigation, 2014, 67 (1): 177- 189. doi: 10.1017/S0373463313000659
6	KRIZHEVSKY A , SUTSKEVER I , HINTON G E . ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60 (6): 1097- 1105.
7	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proc.of the IEEE Conference on Computer Vision and Pattern Recognition, 2014: 580-587.
8	GIRSHICK R. Fast R-CNN[C]//Proc.of the IEEE International Conference on Computer Vision, 2015: 1440-1448.
9	REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[C]//Proc.of the 28th International Conference on Neural Information Processing Systems, 2015: 91-99.
10	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proc.of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 779-788.
11	ADAM V E. You only look twice: rapid multi-scale object detection in satellite imagery[EB/OL].[2019-8-24]. https://arxiv.org/pdf/1805.09512.pdf.
12	JOSEPH R, ALI F. YOLOv3: an incremental Improvement[EB/OL].[2019-8-8]. https://arxiv.org/pdf/1804.02767.pdf.
13	LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proc.of the 14th European Conference on Computer Vision, 2016: 21-37.
14	FU C Y, LIU W, RANGA A, et al. DSSD: deconvolutional single shot detector[EB/OL].[2019-8-23]. https://arxiv.org/pdf/1701.06659.pdf.
15	LIU Y, XU P, ZHANG M H, et al. SAR ship detection using sea-land segmentation-based convolutional neural network[C]//Proc.of the International Workshop on Remote Sensing with Intelligent Processing, 2017.
16	COZZOLINO D, MARTINO D G, POGGI G, et al. A fully convolutional neural network for low-complexity single-stage ship detection in sentinel-1 SAR images[C]//Proc.of the IEEE International Geoscience and Remote Sensing Symposium, 2017: 886-889.
17	ZHAN C, ZHANG L C, ZHONG Z Z, et al. Deep learning approach in automatic iceberg-ship detection with SAR remote sensing data[EB/OL].[2019-10-9]. https://arxiv.org/ftp/arxiv/papers/1812/1812.07367.
18	BELL S, ZITNICK C L, BALA K, et al. Inside-outside net: detecting objects in context with skip pooling and recurrent neural networks[C]//Proc.of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 2874-2883.
19	LIN T Y, DOLLAR P, GIRAHICK R, et al. Feature pyramid networks for object detection[C]//Proc.of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 936-944.
20	YANG X , SUN H , FU K , et al. Automatic ship detection in remote sensing images from google earth of complex scenes based on multiscale rotation dense feature pyramid networks[J]. Remote Sensing, 2018, 10 (1): 132.
21	KANG M , JI K F , LENG X G , et al. Contextual region-based convolutional neural network with multilayer fusion for SAR ship detection[J]. Remote sensing, 2017, 9 (8): 860. doi: 10.3390/rs9080860
22	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc.of the IEEE conference on computer vision and pattern recognition, 2016: 770-778.
23	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]//Proc.of the International Conference on Neural Information Processing System, 2012: 1097-1105.
24	SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL].[2019-10-9].https://arxiv.org/pdf/1409.1556.pdf.
25	SZEGEDY C, LIU W, JIA Y Q, et al. Going deeper with convolution[C]//Proc.of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
26	GAO H, LIU Z, WEINBERGER K Q, et al. Densely connected convolutional networks[EB/OL].[2019-10-9]. https://arxiv.org/pdf/1608.06993.pdf.
27	HU J , SHEN L , ALBANIE S , et al. Squeeze-and-excitation networks[J]. IEEE Trans.on Pattern Analysis and Machine Intelligence, accepted,
28	AN Q Z , PAN Z X , YOU H J . Ship detection in Gaofen-3 SAR images based on sea clutter distribution analysis and deep convolutional neural network[J]. Sensors, 2018, 18 (2): 334. doi: 10.3390/s18020334
29	杜兰, 刘彬, 王燕, 等. 基于卷积神经网络的SAR图像目标检测算法[J]. 电子与信息学报, 2016, 38 (12): 3018- 3025.
	DU L , LIU B , WANG Y , et al. Target detection method based on convolutional neural network for SAR image[J]. Journal of Electronics & Information Technology, 2016, 38 (12): 3018- 3025.
30	ABADI M, AGARWAL A, BARHAM P, et al. Tensorflow: large-scale machine learning on heterogeneous distributed systems[EB/OL].[2019-10-9]. https://arxiv.org/pdf/1603.04467.pdf.
31	李健伟, 曲长文, 彭书娟, 等. 基于卷积神经网络的SAR图像舰船目标检测[J]. 系统工程与电子技术, 2018, 40 (9): 1953- 1959.
	LI J W , QU C W , PENG S J , et al. Ship detection in SAR images based on convolutional neural network[J]. Systems Engineering and Electronics, 2018, 40 (9): 1953- 1959.
32	杨龙, 苏娟, 李响. 基于深度卷积神经网络的SAR舰船目标检测[J]. 系统工程与电子技术, 2019, 41 (9): 1990- 1997.
	YANG L , SU J , LI X . Ship detection in SAR images based on deep convolutional neural network[J]. Systems Engineering and Electronics, 2019, 41 (9): 1990- 1997.

数据集	星载SAR图像	子图像		所含目标数
数据集	星载SAR图像	含目标子图像	无目标子图像	所含目标数
训练集	5	1 936	-	4 284
测试集	2	591	2 945	1 232

β取值	N_{total_target}	N_td	N_fd	P_d/%	P_f/%	F₁
1	1 270	1 101	169	89.37	13.31	0.880
2	1 284	1 124	160	91.23	12.46	0.893
3	1 289	1 135	154	92.13	11.95	0.900
4	1 312	1 153	159	93.59	12.12	0.906
5	1 287	1 161	126	94.24	9.79	0.922
6	1 271	1 074	197	87.18	15.50	0.858
7	1 239	1 077	162	87.42	13.07	0.872

子测试集	舰船尺度范围	样本量	所含目标数	P_d/%	P_f/%	F₁
1	≥48²	90	197	92.89	10.29	0.913
2	24²~48²	220	462	95.02	9.67	0.926
3	≤24²	120	268	93.66	10.36	0.916

方法	N_{total_target}	N_td	N_fd	P_d/%	P_f/%	F₁	测试耗时/s
FPN	1 329	1 087	242	88.23	18.21	0.849	0.28
Faster R-CNN+	1 316	1 083	233	87.91	17.71	0.850	0.22
CMF-RCNN	1 272	1 075	197	87.26	15.49	0.859	0.22
SSD	1 345	1 124	221	91.23	16.43	0.872	0.20
SSD+	1 353	1 147	206	93.10	15.23	0.887	0.21
DFPN	1 275	1 116	159	90.58	12.47	0.890	0.30
本文算法	1 287	1 161	126	94.24	9.79	0.922	0.35

[1]	宋爽, 张悦, 张琳娜, 岑翼刚, 李浥东. 基于深度学习的轻量化目标检测算法[J]. 系统工程与电子技术, 2022, 44(9): 2716-2725.
[2]	苗添, 曾虹程, 王贺, 陈杰. 基于迭代阈值分割的星载SAR洪水区域快速提取[J]. 系统工程与电子技术, 2022, 44(9): 2760-2768.
[3]	王彩云, 吴钇达, 王佳宁, 马璐, 赵焕玥. 基于改进的CNN和数据增强的SAR目标识别[J]. 系统工程与电子技术, 2022, 44(8): 2483-2487.
[4]	傅东宁, 廖桂生, 黄岩, 张邦杰, 王幸. 基于图拉普拉斯嵌入的合成孔径雷达时变窄带干扰抑制算法[J]. 系统工程与电子技术, 2022, 44(6): 1846-1853.
[5]	盖明慧, 张苏, 孙卫天, 倪育德, 杨磊. 复数兼容全变分SAR目标结构特征增强[J]. 系统工程与电子技术, 2022, 44(6): 1862-1872.
[6]	徐安林, 张毓, 周峰. 基于Beta过程的高分辨ISAR成像[J]. 系统工程与电子技术, 2022, 44(6): 1873-1879.
[7]	韦娟, 杨皇卫, 宁方立. 基于NMF与CNN联合优化的声学场景分类[J]. 系统工程与电子技术, 2022, 44(5): 1433-1438.
[8]	纪朋徽, 代大海, 邢世其, 冯德军. 密集虚假运动目标生成方法[J]. 系统工程与电子技术, 2022, 44(5): 1502-1511.
[9]	刘丰恺, 黄大荣, 郭新荣, 冯存前. 基于吕氏分布的机动目标参数化平动补偿方法[J]. 系统工程与电子技术, 2022, 44(4): 1166-1173.
[10]	陈冬, 句彦伟. 基于语义分割实现的SAR图像舰船目标检测[J]. 系统工程与电子技术, 2022, 44(4): 1195-1201.
[11]	周晓玲, 张朝霞, 鲁雅, 王倩, 王琨琨. 基于改进R-FCN的SAR图像识别[J]. 系统工程与电子技术, 2022, 44(4): 1202-1209.
[12]	方伟, 王玉, 闫文君, 林冲. 基于神经网络的符号化飞行动作识别[J]. 系统工程与电子技术, 2022, 44(3): 737-745.
[13]	孙晶明, 虞盛康, 孙俊. 基于深度学习的HRRP识别姿态敏感性分析[J]. 系统工程与电子技术, 2022, 44(3): 802-807.
[14]	杨磊, 张苏, 盖明慧, 方澄. 高分辨SAR目标成像方向性结构特征增强[J]. 系统工程与电子技术, 2022, 44(3): 808-818.
[15]	刘恒燕, 张立民, 闫文君, 钟兆根, 凌青, 梁晓军. 基于WBP-CNN算法的LDPC译码[J]. 系统工程与电子技术, 2022, 44(3): 1030-1035.