首页 > 论文 > 液晶与显示 > 33卷 > 6期(pp:497-503)

基于统计信息的改进滑动平均目标检测算法

Improved sliding average target detection algorithm based on statistical information

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

随着实时视频监控技术的发展,运动目标的检测在众多领域都有着广泛的应用。针对当下运动目标检测算法易受环境噪声和光照突变的影响,为了提高算法的稳定性和自主检测能力,本文系统研究了背景减除法和帧差法的不足和优势,提出了一种基于统计信息的改进滑动平均算法。首先,依据帧间差分对光照不敏的特性,粗略分离出背景和前景部分,同时去除掉其他随机孤立噪声,然后在滑动平均算法中引入对前景点的统计计数,自适应更新背景建模学习速率,加速背景模型收敛,最后将滑动平均与三帧差分两者形态学处理后的目标团块进行与运算,得到准确的前景区域。实验结果表明,本算法可以在树叶晃动和光照变化中精确定位运动目标,正确前景点率平均提升7.4%,得益于帧差法特性,误检前景点率显著降低,鲁棒性和自适应性更好。

Abstract

With the development of real-time video surveillance, moving target detection algorithm has been used in many fields. The current moving object detection algorithm is most vulnerable to environmental noise and light mutation. In order to improve the stability and self-adaptability of the algorithm, an improved moving average target detection algorithm based on statistical information is proposed. Firstly, on the basis of the characteristics of inter-frame difference that is insensitive about light, the background and foreground parts are separated and other random noise is also removed. Secondly, the information of statistical counting of foreground points is introduced in the moving average algorithm to adaptive update background modeling learning rate which is effective to accelerate the convergence of the background model. Finally, to get accurate moving target position we should process the results of the two algorithms by further using morphological methods and AND operation. Experimental results show that the proposed algorithm can accurately locate moving targets under the condition of light mutation and leaves shaking. Compared to the previous algorithm, the TP rate increases by 7.4% and the FP rate significantly decreases. The proposed algorithm has better robustness and adaptability.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:TP391

DOI:10.3788/yjyxs20183306.0497

所属栏目:图像处理

基金项目:国家自然科学基金青年科学基金项目(No.61602432)

收稿日期:2017-12-26

修改稿日期:2018-03-27

网络出版日期:--

作者单位    点击查看

李大维:中国科学院 长春光学精密机械与物理研究所,吉林 长春 130033中国科学院大学,北京 100049
孙海江:中国科学院 长春光学精密机械与物理研究所,吉林 长春 130033
刘伟宁:中国科学院 长春光学精密机械与物理研究所,吉林 长春 130033
刘培勋:中国科学院 长春光学精密机械与物理研究所,吉林 长春 130033

联系人作者:李大维(lidawei315@mails.ucas.ac.cn)

备注:李大维(1997-),男,山西忻州人,硕士研究生,2015年于西安电子科技大学获得学士学位,主要从事数字图像处理,目标检测与识别技术研究。

【1】黄凯奇,陈晓棠,康运锋,等.智能视频监控技术综述[J].计算机学报,2015,38(6):1093-1118.
HUANG K Q, CHEN X T, KANG Y F, et al. Intelligent visual surveillance: a review [J]. Chinese Journal of Computers, 2015, 38(6): 1093-1118. (in Chinese)

【2】骆云志,刘治红.视频监控技术发展综述[J].兵工自动化,2009,28(1):1-3, 11.
LUO Y Z, LIU Z H. Review of video monitoring technology development [J]. Ordnance Industry Automation, 2009, 28(1): 1-3, 11. (in Chinese)

【3】ZOU J H, ZHAO Q C. Video object segmentation using background model based on pixel time series clustering [C]//Proceedings of the 23rd International Conference on Automation and Computing (ICAC). Huddersfield, UK: IEEE, 2017: 1-6.

【4】LIU C G, WANG S. An adapting to light change pixel layer based background model for moving objects detection in a dynamic scene [C]//Proceedings of the 11th International Symposium on Distributed Computing and Applications to Business, Engineering & Science (DCABES). Guilin, China: IEEE, 2012: 405-408.

【5】RASHID M E, THOMAS V. A background foreground competitive model for background subtraction in dynamic background [J]. Procedia Technology, 2016, 25: 536-543.

【6】ALI I, MILLE J, TOUGNE L. Adding a rigid motion model to foreground detection: application to moving object detection in rivers [J]. Pattern Analysis and Applications, 2014, 17(3): 567-585.

【7】吴桐,王玲.基于帧差分块的混合高斯背景模型[J].计算机工程与应用,2014,50(23):176-180.
WU T, WANG L. Gaussian mixture background model based on inter-frame differencing blocks [J]. Computer Engineering and Applications, 2014, 50(23): 176-180. (in Chinese)

【8】刘翔,周桢.基于分块背景建模的运动目标检测技术[J].液晶与显示,2011,26(6):831-835.
LIU X, ZHOU Z. Moving objectives detecting technology based on block processing of background model establishment [J]. Chinese Journal of Liquid Crystals and Displays, 2011, 26(6): 831-835. (in Chinese)

【9】闵卫东,郭晓光,韩清.改进的ViBe算法及其在交通视频处理中的应用[J].光学 精密工程,2017,25(3):806-811.
MIN W D, GUO X G, HAN Q. An improved ViBe algorithm and its application in traffic video processing [J]. Optics and Precision Engineering, 2017, 25(3): 806-811. (in Chinese)

【10】万缨,韩毅,卢汉清.运动目标检测算法的探讨[J].计算机仿真,2006,23(10):221-226.
WAN Y, HAN Y, LU H Q. The methods for moving object detection [J]. Computer Simulation, 2006, 23(10): 221-226. (in Chinese)

【11】ZHU W, TIAN L, DI F, et al. Dynamic targets detection for robotic applications using panoramic camera based on optical flow [J]. Applied Mechanics and Materials, 2013, 376: 455-460.

【12】刘洪彬,常发亮.权重系数自适应光流法运动目标检测[J].光学 精密工程,2016,24(2):460-468.
LIU H B, CHANG F L. Moving object detection by optical flow method based on adaptive weight coefficient [J]. Optics and Precision Engineering, 2016, 24(2): 460-468. (in Chinese)

【13】康晓晶,吴谨.基于高斯背景建模的目标检测技术[J].液晶与显示,2010,25(3):454-459.
KANG X J, WU J. Object Detecting technology based on gauss background modeling [J]. Chinese Journal of Liquid Crystals and Displays, 2010, 25(3): 454-459. (in Chinese)

【14】赵建.基于三帧差法的运动目标检测方法研究[D].西安:西安电子科技大学,2013.
ZHAO J. The research of moving target detection method based on three-frame difference [D]. Xi’an: Xidian University, 2013. (in Chinese)

【15】何林宣.基于视频序列的双背景建模算法[D].长春:吉林大学,2014.
HE L X. Algorithm on video-based dual background modeling [D]. Changchun: Jilin University, 2014. (in Chinese)

引用该论文

LI Da-wei,SUN Hai-jiang,LIU Wei-ning,LIU Pei-xun. Improved sliding average target detection algorithm based on statistical information[J]. Chinese Journal of Liquid Crystals and Displays, 2018, 33(6): 497-503

李大维,孙海江,刘伟宁,刘培勋. 基于统计信息的改进滑动平均目标检测算法[J]. 液晶与显示, 2018, 33(6): 497-503

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF