• 《工程索引》(EI)刊源期刊
  • 综合性科学技术类中文核心期刊
  • 中国科技论文统计源期刊
  • 中国科学引文数据库来源期刊

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

复杂环境下一种基于SiamMask的时空预测移动目标跟踪算法

周珂 张浩博 付冬梅 赵志毅 曾惠

周珂, 张浩博, 付冬梅, 赵志毅, 曾惠. 复杂环境下一种基于SiamMask的时空预测移动目标跟踪算法[J]. 工程科学学报, 2020, 42(3): 381-389. doi: 10.13374/j.issn2095-9389.2019.06.06.005
引用本文: 周珂, 张浩博, 付冬梅, 赵志毅, 曾惠. 复杂环境下一种基于SiamMask的时空预测移动目标跟踪算法[J]. 工程科学学报, 2020, 42(3): 381-389. doi: 10.13374/j.issn2095-9389.2019.06.06.005
ZHOU Ke, ZHANG Hao-bo, FU Dong-mei, ZHAO Zhi-yi, ZENG Hui. Design and implementation of multi-feature fusion moving target detection algorithms in a complex environment based on SiamMask[J]. Chinese Journal of Engineering, 2020, 42(3): 381-389. doi: 10.13374/j.issn2095-9389.2019.06.06.005
Citation: ZHOU Ke, ZHANG Hao-bo, FU Dong-mei, ZHAO Zhi-yi, ZENG Hui. Design and implementation of multi-feature fusion moving target detection algorithms in a complex environment based on SiamMask[J]. Chinese Journal of Engineering, 2020, 42(3): 381-389. doi: 10.13374/j.issn2095-9389.2019.06.06.005

复杂环境下一种基于SiamMask的时空预测移动目标跟踪算法

doi: 10.13374/j.issn2095-9389.2019.06.06.005
基金项目: 国家自然科学基金资助项目(61375010);北京科技大学基本科研业务费资助项目(FRF-OT-18-020SY)
详细信息
    通讯作者:

    E-mail: cocofay126@126.com

  • 中图分类号: TG142.71

Design and implementation of multi-feature fusion moving target detection algorithms in a complex environment based on SiamMask

More Information
  • 摘要: 随着无人工厂、智能安监等技术在制造业领域的深入应用,以视觉识别预警系统为代表的复杂环境下动态识别技术成为智能工业领域的重要研究内容之一。在本文所述的工业级视觉识别预警系统中,操作人员头发区域由于其具有移动形态非规则性、运动无规律性的特点,在动态图像中的实时分割较为困难。针对此问题,提出一种基于SiamMask模型的时空预测移动目标跟踪算法。该算法将基于PyTorch深度学习框架的SiamMask单目标跟踪算法与ROI检测及STC时空上下文预测算法相融合,根据目标时空关系的在线学习,预测新的目标位置并对SiamMask模型进行算法校正,实现视频序列中的目标快速识别。实验结果表明,所提出的算法能够克服环境干扰、目标遮挡对跟踪效果的影响,将目标跟踪误识别率降低至0.156%。该算法计算时间成本为每秒30帧,比改进前的SiamMask模型帧率每秒提高3.2帧,算法效率提高11.94%。该算法达到视觉识别预警系统准确性、实时性的要求,对移动目标识别算法模型的复杂环境应用具有借鉴意义。
  • 图  1  SiamMask模型算法流程图[5]. (a)三分支变型架构;(b)二分支变型架构核心

    Figure  1.  SiamMask model algorithmic flow chart[5]: (a) three-branch variant architecture; (b) two-branch variant head

    图  2  SiamMask模型面部检测效果. (a)束发头部跟踪;(b)长发头部跟踪Ⅰ;(c)长发头部跟踪Ⅱ

    Figure  2.  SiamMask model face detection effect: (a) bundle head tracking; (b) long hair head tracking I; (c) long hair head tracking II

    图  3  SiamMask模型测试误识别现象. (a)深色干扰源误识别;(b)肉色干扰误识别;(c)头发遮挡误识别

    Figure  3.  SiamMask model test misrecognition phenomenon: (a) misidentification of dark interference sources; (b) misidentification of flesh color interference; (c) misidentification of hair occlusion

    图  4  基于SiamMask模型的时空预测移动目标跟踪算法框架图

    Figure  4.  Framework of spatiotemporal prediction moving target tracking algorithms based on the SiamMask Model

    图  5  车工监控视频ROI提取结果. (a)原始画面;(b)ROI提取画面

    Figure  5.  ROI extraction result of a locomotive monitoring video: (a) original picture; (b) ROI extraction picture

    图  6  运动图像检测灰度图. (a)无运动目标时的原图/灰度图;(b)运动目标出现时的原图/灰度图

    Figure  6.  Gray level image of moving image detection: (a) original image / gray level image without moving object; (b) original image / gray level image when moving object appears

    图  7  图像灰度及阈值化处理. (a)原始视频图像;(b) 灰度化处理结果;(c) 阈值化处理结果

    Figure  7.  Gray level and threshold processing of image: (a) original video image; (b) grayscale processing results; (c) threshold processing results

    图  8  算法训练/测试准确率及损失率曲线. (a)训练集准确率曲线图;(b)训练集损失率曲线图;(c)测试集准确率曲线图;(d)测试集损失率曲线图

    Figure  8.  Algorithm training/test accuracy and loss rate curve: (a) training set accuracy curve; (b) training set loss curve; (c) test set accuracy curve; (d) test set loss curve

    图  9  固定危险区划分示意图. (a)视频危险区划分图;(b)危险级别划分示意图

    Figure  9.  Fixed danger zone division diagram: (a) video dangerous zone division map; (b) diagram of hazard classification

    图  10  头发目标跟踪报警结果. (a)第30帧;(b)第35帧;(c)第60帧;(d)第70帧;(e)第75帧;(f)第80帧

    Figure  10.  Hair target tracking alarm results: (a) frame 30; (b) frame 35; (c) frame 60; (d) frame 70; (e) frame 75; (f) frame 80

    表  1  SiamMask模型目标跟踪效果统计

    Table  1.   Statistics of target tracking effect of the SiamMask model

    Video No.Frame number of false detectionAnalysis on the causes of false inspectionTotal frames Failure rate/%
    10Little change in this movement3610
    287Misidentified as dark cloth28830.21
    398Part of the face is blocked by the hair19251.04
    4674Initialization offset, screen will pop up in recognition138048.84
    5131The target moves out of the screen slightly and the recognition is lost24054.58
    6753Large proportion of face selection in initialization area136055.37
    70Accurate initialization and small action range2410
    下载: 导出CSV

    表  2  基于SiamMask模型的时空预测算法目标跟踪效果统计

    Table  2.   Statistics of the target tracking effect of the spatiotemporal prediction algorithms based on the SiamMask model

    Video No.Frame number of false detection Analysis on the causes of false inspectionTotal frames Failure rate/%
    10Little change in this movement3610
    20Misidentified as dark cloth2880
    31Part of the face is blocked by the hair1920.52
    42Initialization offset, screen will pop up in recognition13800.15
    51The target moves out of the screen slightly and the recognition is lost2400.42
    60Large proportion of face selection in initialization area13600
    70Accurate initialization and small action range2410
    下载: 导出CSV
  • [1] Xing J L, Ai H Z, Lao S H. Multiple human tracking based on multi-view upper-body detection and discriminative learning // 2010 20th International Conference on Pattern Recognition. Istanbul, 2010: 1698
    [2] Liu L W, Xing J L, Ai H Z, et al. Hand posture recognition using finger geometric feature // Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012). Tsukuba, 2012: 565
    [3] Bertinetto L, Valmadre J, Henriques J F, et al. Fully-convolutional siamese networks for object tracking // European Conference on Computer Vision. Cham: Springer, 2016: 850
    [4] Li B, Yan J J, Wu W, et al. High performance visual tracking with siamese region proposal network // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018: 8971
    [5] Wang Q, Zhang L, Bertinetto L, et al. Fast online object tracking and segmentation: a unifying approach // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2019: 1328
    [6] 李江昀, 赵义凯, 薛卓尔, 等. 深度神经网络模型压缩综述. 工程科学学报, 2019, 41(10): 1229

    Li J Y, Zhao Y K, Xue Z E, et al. A survey of model compression for deep neural networks. Chin J Eng, 2019, 41(10): 1229
    [7] 沈丽丽, 杭宁. 联合多种边缘检测算子的无参考质量评价算法. 工程科学学报, 2018, 40(8): 996

    Shen L L, Hang N. No-reference image quality assessment using joint multiple edge detection. Chin J Eng, 2018, 40(8): 996
    [8] 王鑫, 王振民. 一种改进的基于Camshift的粒子滤波实时目标跟踪算法. 中国图像图形学报, 2010, 15(10):1507

    Wang X, Tang Z M. An improved camshaft-based particle filter algorithm for real-time target tracking. J Image Graph, 2010, 15(10): 1507
    [9] 初红霞, 谢忠玉, 王科俊. 一种结合颜色纹理直方图的改进型Camshift目标跟踪算法. 西安交通大学学报, 2018, 52(3):145

    Chu H X, Xie Z Y, Wang K J. An improved camshift target tracking algorithm based on joint color-texture histogram. J Xian Jiaotong Univ, 2018, 52(3): 145
    [10] Yacoob Y, Davis L S. Detection and analysis of hair. IEEE Trans Pattern Anal Mach Intell, 2006, 28(7): 1164 doi: 10.1109/TPAMI.2006.139
    [11] 傅文林, 胡福乔. 基于mean shift的头发自动检测. 微型电脑应用, 2010, 26(9):62

    Fu W L, Hu F Q. Hair detection with mean shift. Microcomput Appl, 2010, 26(9): 62
    [12] 王志一, 杨大利. 人脸识别中发型遮挡检测方法研究. 微型机与应用, 2016, 35(2):32 doi: 10.3969/j.issn.1674-7720.2016.02.012

    Wang Z Y, Yang D L. Research on detection method of hair-occlusion in face recognition. Microcomput Its Appl, 2016, 35(2): 32 doi: 10.3969/j.issn.1674-7720.2016.02.012
    [13] 丁承君, 闫彬. 时空上下文与CamShift相结合的目标跟踪算法. 传感器与微系统, 2018, 37(5):108

    Ding C J, Yan B. Target tracking algorithm combines STC with CamShift. Transduc Microsyst Technol, 2018, 37(5): 108
    [14] Eriksson D, Bindel D, Shoemaker C. Surrogate Optimization Toolbox Hithub (2019-1-5)[2019-6-5]. https://github.com/dme65/pySOT
    [15] Qiang Wang, Li Zhang, Luca Bertinetto, et al. Torr. SiamMask Hithub (2019-5-5)[2019-6-5]. https://github.com/foolwood/SiamMask
    [16] Zhang L, Suganthan P N. Robust visual tracking via co-trained kernelized correlation filters. Pattern Recognit, 2017, 69: 82 doi: 10.1016/j.patcog.2017.04.004
    [17] Chen C, Wu Z X, Jiang Y G. Emotion in context: Deep semantic feature fusion for video emotion recognition // Proceedings of the 24th ACM International Conference on Multimedia. New York, 2016: 127
    [18] 黄煜, 黄翔, 李泷杲, 等. 存在视场丢失的机器视觉精度补偿方法, 工程科学学报, 2018, 40(3): 381

    Huang Y, Huang X, Li S G, et al. Compensation method for a robot vision system with an occluded camera field. Chin J Eng, 2018, 40(3): 381
    [19] 张晶, 王旭, 范洪博. 自适应学习的时空上下文目标跟踪算法. 计算机工程, 2018, 44(6):294 doi: 10.3969/j.issn.1000-3428.2018.06.050

    Zhang J, Wang X, Fan H B. Spatio temporal context target tracking algorithm of self-adaption learning. Comput Eng, 2018, 44(6): 294 doi: 10.3969/j.issn.1000-3428.2018.06.050
    [20] Pont-Tuset J, Perazzi F, Caelles S, et al. The 2017 Davis Challenge on Video Object Segmentation (2017-4-3)[2019-6-5]. https://arxiv.org/abs/1704.00675
  • 加载中
图(10) / 表(2)
计量
  • 文章访问数:  859
  • HTML全文浏览量:  2219
  • PDF下载量:  32
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-06-06
  • 刊出日期:  2020-03-01

目录

    /

    返回文章
    返回