结合深度轮廓特征的改进孪生网络跟踪算法

doi:10.19665/j.issn1001-2400.2020.03.006

摘要/Abstract

摘要：

针对现有的孪生网络目标跟踪算法存在跟踪漂移的问题,提出了一种结合深度轮廓生成网络的改进孪生网络跟踪模型,以实现复杂背景下对任何目标的稳定检测与跟踪。首先,轮廓检测网络自动获取目标的封闭轮廓信息,并利用泛洪聚类算法获得轮廓模板;然后将轮廓模板与搜索区域输入到改进的孪生网络,获得最优跟踪评分值,并自适应地更新轮廓模板。若目标被遮挡或跟踪丢失,则采用检测网络全视场搜索目标,实现全过程稳定跟踪。大量定性及定量仿真试验结果表明,这种改进模型不仅能够提高复杂背景下目标的跟踪性能,还能提升机载系统的反应时间,适合于工程应用。

关键词: 目标跟踪, 深度学习, 孪生网络, 轮廓检测网络, 目标检测, 自适应模板更新

Abstract:

The existing Siamese object tracking algorithms easily lead to tracking drift under the influence of object deformation and occlusion, this paper proposes an improved object tracking algorithm based on deep contour extraction networks to achieve stable detection and tracking of any object under complex backgrounds. First, the contour detection network automatically obtains the closed contour information on the object and uses the flood-filling clustering algorithm to obtain the contour template. Then, the contour template and the search area are input into the improved Siamese network so as to obtain the optimal tracking score value and adaptively update the contour template. If the object is fully obscured or lost, the Yolov3 network is used to search the object in the entire field of view to achieve stable tracking throughout the process. A large number of qualitative and quantitative simulation results show that the improved model can not only improve the object tracking performance under complex backgrounds, but also improve the response time of airborne systems, which is suitable for engineering applications.

Key words: object tracking, deep learning, Siamese network, contour detection network, object detection, adaptive template updating

中图分类号:

TP183

余志超,张瑞红. 结合深度轮廓特征的改进孪生网络跟踪算法[J]. 西安电子科技大学学报, 2020, 47(3): 40-49.

YU Zhichao,ZHANG Ruihong. Improved Siamese network based object tracking combined with the deep contour feature[J]. Journal of Xidian University, 2020, 47(3): 40-49.

图/表 9

图1

图2

图3

表1

表2

图4

图5

图6

图7

参考文献 22

[1]	史光普, 李明月. 机载光电探测设备对地定位算法及其仿真分析[J]. 火力与指挥控制, 2019,44(7):178-181.
	SHI Guangpu, LI Mingyue. New Algorithm and Simulation Analysis for Ground Target Location with Airborne Photoelectric Detection Equipment[J]. Fire Control & Command Control, 2019,44(7):178-181.
[2]	宋建锋, 苗启广, 申猛, 等. 多特征融合的相关滤波红外单目标跟踪算法[J]. 西安电子科技大学学报, 2019,46(5):142-147.
	SONG Jianfeng, MIAO Qiguang, SHEN Meng, et al. Algorithm for Tracking an Infrared Single Target Based on Correlation Filtering with Multi-feature Fusion[J]. Journal of Xidian University, 2019,46(5):142-147.
[3]	王洪雁, 邱贺磊, 裴腾达. 利用判别字典学习的视觉跟踪方法[J]. 西安电子科技大学学报, 2019,46(4):150-158.
	WANG Hongyan, QIU Helei, PEI Tengda. Visual Tracking Method Using Discriminant Dictionary Learning[J]. Journal of Xidian University, 2019,46(4):150-158.
[4]	王树亮, 毕大平, 阮怀林. 机动目标跟踪波形和检测门限自适应算法[J]. 西安电子科技大学学报, 2019,46(3):96-101.
	WANG Shuliang, BI Daping, RUAN Huailin. Waveform and Detection Threshold Self-adaption Algorithm for Maneuvering Target Tracking[J]. Journal of Xidian University, 2019,46(3):96-101.
[5]	GALTERI L, SEIDENARI L, BERTINI M, et al. Spatio-temporal Closed-loop Object Detection[J]. IEEE Transactions on Image Processing, 2017,12(9):1-13. doi: 10.1109/TIP.2002.806244
[6]	孙护军. 基于双模式遮挡检测机制的红外目标跟踪算法[J]. 红外技术, 2018,40(9):902-907.
	SUN Hujun. An Infrared Object Tracking Algorithm Based on Dual-mode Context Occlusion Detection Mechanism[J]. Infrared Technology, 2018,40(9):902-907.
[7]	XU Z, LUO H, HUI B, et al. Siamese Tracking from Single Point Initialization[J]. Sensors, 2019,19(3):514. doi: 10.3390/s19030514
[8]	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet Classification with Deep Convolutional Neural Networks [C]// Proceedings of the 2012 25th International Conference on Neural Information Processing Systems: 1. Vancouver: NIPS, 2012: 1097-1105.
[9]	WANG N Y, YEUNG D Y. Learning a Deep Compact Image Representation for Visual Tracking [C]//Proceedings of the 2013 26th International Conference on Neural Information Processing Systems: 1. Vancouver: NIPS, 2013: 809-817.
[10]	WANG L, OUYANG W, WANG X, et al. Visual Tracking with Fully Convolutional Networks [C]//Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2015: 3119-3127.
[11]	XU Z, LUO H, HUI B, et al. Siamese Tracking with Adaptive Template-updating Strategy[J]. Applied Sciences, 2019,9(18):3725. doi: 10.3390/app9183725
[12]	ZHU Z, WANG Q, LI B, et al. Distractor-aware Siamese Networks for Visual Object Tracking [C]//Lecture Notes in Computer Science: 11213 LNCS. Heidelberg : Springer Verlag, 2018: 103-119.
[13]	YAN B, ZHAO H, WANG D, et al. 'Skimming-Perusal' Tracking: a Framework for Real-time and Robust Long-term Tracking [C]//Proceedings of the 2019 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2019: 2385-2393.
[14]	ZHANG Y, WANG L, QI J, et al. Structured Siamese Network for Real-time Visual Tracking [C]//Lecture Notes in Computer Science: 11213 LNCS. Heidelberg : Springer Verlag, 2018: 355-370.
[15]	ZHANG Z, PENG H. Deeper and Wider Siamese Networks for Real-time Visual Tracking [C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2019: 4586-4595.
[16]	HE Y, HU T, ZENG D. Scan-flood Fill (SCAFF): an Efficient Automatic Precise Region Filling Algorithm for Complicated Regions[CP/OL]. [2019-07-20].https://www.researchgate.net/publication/333678982.
[17]	REDMON J, FARHADI A. Yolov3: an Incremental Improvement[J/OL]. [2019-07-20].https://arxiv.org/pdf/1804.02767.
[18]	BERTINETTO L, VALMADRE J, HENRIQUES J F, et al. Fully-convolutional Siamese Networks for Object Tracking[J]. Lecture Notes in Computer Science : 9914 LNCS. Heidelberg : Springer Verlag, 2016: 850-865.
[19]	HELD D, THRUN S, SAVARESE S. Learning to Track at 100fps with Deep Regression Networks[CP/OL]. [2019-07-20].https://arxiv.org/pdf/1604.01802.pdf.
[20]	LUKEZIC A, VOJIR T, ZAJC L C, et al. Discriminative Correlation Filter with Channel and Spatial Reliability [C]//Proceedings of the 2017 30th IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway: IEEE, 2017: 4847-4856.
[21]	TAO R, GAVVES E, SMEULDERS A W M. Siamese Instance Search for Tracking [C]//Proceedings of the 2017 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. Washington: IEEE Computer Society, 2017: 1420-1429.
[22]	LI B, YAN J, WU W, et al. High Performance Visual Tracking with Siamese Region Proposal Network [C]//Proceedings of the 2018 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2018: 8971-8980.

参数	C_{1_2}	P₁	C_{2_2}	P₂	C_{3_3}	P₃	C_{4_3}	P₄	C_{5_3}
感受野	5	6	14	16	40	44	92	100	196
步幅	1	2	2	4	4	8	8	16	16

参数	C₁	P₁	C₂	P₂	C₃	C₄	C₅
感受野	11	15	31	39	71	103	135
步幅	4	8	8	16	16	16	16