铁路客运站复杂环境中的运动目标检测

孙首群; 刘康亚; 刘硕妍; 吕晓军; 詹璇

doi:10.19818/j.cnki.1671-1637.2013.03.016

铁路客运站复杂环境中的运动目标检测

doi: 10.19818/j.cnki.1671-1637.2013.03.016

1.
上海理工大学机械工程学院, 上海 200093
2.
中国铁道科学研究院电子计算技术研究所, 北京 100081

基金项目:

国家863计划项目 2009AA11211

铁道部专项资金支持课题研究项目 J2011X007

上海市教委重点学科建设项目 J50503

详细信息

作者简介:
孙首群(1964-), 男, 河南郑州人, 上海理工大学副教授, 工学博士, 从事机电系统的电热耦合和系统动力学研究

中图分类号: U491.116
计量
- 文章访问数: 857
- HTML全文浏览量: 192
- PDF下载量: 729
- 被引次数: 0
出版历程
- 收稿日期: 2013-01-18
- 刊出日期: 2013-06-25

Moving target detection in complex environment of railway station

1.
School of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
2.
Institute of Computing Technology, China Academy of Railway Sciences, Beijing 100081, China

More Information

Author Bio:
SUN Shourqun(1964-), male, associate professor, PhD, +86-21-55275056, jrssq@163.com

Article Text (Baidu Translation)

摘要

摘要: 采用分层组织的形式将传统高斯混合模型分为背景层、竞争层和噪声层, 各层分别采用不同的更新机制, 在各层之间引入晋级和降级机制以纠正可能存在的误判。采用基于轮廓检测的噪声滤波实现噪声层更新以消除噪声, 并利用直方图匹配检测伪前景区域以提高对背景变化的适应能力。使用停车场视频和铁路客运站候车室视频对改进后高斯混合模型的检测效果进行了验证。验证结果表明: 改进的高斯混合模型有效避免了长期静止的目标被融入背景, 降低了光线突变或摄像机噪声的干扰, 加快了背景改变时模型的更新速度, 目标检测速度比传统GMM提高了10%。检则方法满足了铁路客运站智能视频监控实时性和准确性的要求, 为视频分析奠定了基础。
- 运动目标检测 /
- 高斯混合模型 /
- 直方图匹配 /
- 噪声滤波 /
- 分层组织 /
- 轮廓检测
Abstract: Traditional GMM(Gaussian mixture model) was dived into background layer, completion layer and noise layer by using hierarchical organization.Diverse update mechanisms were applied in different layers. In order to correct possible misjudgment, promotion and downgraded mechanisms were introduced between layers. To eliminate noise, noise layer was updated by using noise filter based on contour detection. In order to improve the adaptability for changing background, pseudo foreground area was detected by using histogram matching. The detection effect of improved GMM was verified by using the videos of station and parking lot. Verification result indicates that the problem of long-term static target being merged into background is settled. The impact of light mutations or camera noise is reduced. The updating speed of model increases when the background changes. Detection speed increases by 10% compared with traditional GMM.The efficiency and accuracy of moving target detection in railway station are improved by improved GMM, and the foundation for intelligent video analysis is laid.
- moving target detection /
- Gaussian mixture model /
- histogram matching /
- noise filter /
- hierarchical organization /
- contour detection

HTML全文

图 1 改进的GMM算法

Figure 1. Improved GMM algorithm

下载: 全尺寸图片幻灯片

图 2 第150帧停车场图像中运动目标检测结果

Figure 2. Detection results of moving objects in image 150 of parking lot

下载: 全尺寸图片幻灯片

图 3 第500帧停车场图像中运动目标检测结果

Figure 3. Detection results of moving objects in image 500 of parking lot

下载: 全尺寸图片幻灯片

图 4 第800帧停车场图像中运动目标检测结果

Figure 4. Detection results of moving objects in image 800 of parking lot

下载: 全尺寸图片幻灯片

图 5 第1 200帧停车场图像中运动目标检测结果

Figure 5. Detection results of moving objects in image 1 200 of parking lot

下载: 全尺寸图片幻灯片

图 6 第300帧候车室图像中运动目标检测结果

Figure 6. Detection results of moving objects in image 300 of waiting room

下载: 全尺寸图片幻灯片

图 7 第800帧候车室图像中运动目标检测结果

Figure 7. Detection results of moving objects in image 800 of waiting room

下载: 全尺寸图片幻灯片

图 8 第1 200帧候车室图像中运动目标检测结果

Figure 8. Detection results of moving objects in image 1 200 of waiting room

下载: 全尺寸图片幻灯片

图 9 第2 500帧候车室图像中运动目标检测结果

Figure 9. Detection results of moving objects in image 2 500 of waiting room

下载: 全尺寸图片幻灯片

图 10 第3 500帧候车室图像中运动目标检测结果

Figure 10. Detection results of moving objects in image 3 500 of waiting room

下载: 全尺寸图片幻灯片

表 1 耗时对比

Table 1. Comparison of consuming times ms·帧^-1

视频	方法	最大耗时	最小耗时	平均耗时
停车场	传统GMM	328	296	313
	改进GMM	312	265	281
	传统GMM	468	156	235
	改进GMM	406	171	203

下载: 导出CSV

参考文献(18)

[1]	郑锦, 李波. 视频序列中运动对象检测技术的研究现状与展望[J]. 计算机应用研究, 2008, 25(12): 3534-3540. doi: 10.3969/j.issn.1001-3695.2008.12.004 ZHENG Jin, LI Bo. Prospects and current studies on motion object detection in video sequences[J]. Application Research of Computers, 2008, 25(12): 3534-3540. (in Chinese). doi: 10.3969/j.issn.1001-3695.2008.12.004
[2]	STAUFFER C, GRIMSON W E L. Learning patterns of activity using real-time tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(8): 747-757. doi: 10.1109/34.868677
[3]	杨国亮, 王志良, 牟世堂, 等. 一种改进的光流算法[J]. 计算机工程, 2006, 32(15): 187-188, 226. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC200615065.htm YANG Guo-liang, WANG Zhi-liang, MU Shi-tang, et al. An improved optical flow algorithm[J]. Computer Engineering, 2006, 32(15): 187-188, 226. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC200615065.htm
[4]	娄路, 赵玲, 耿涛. 运动车辆检测与跟踪方法[J]. 交通运输工程学报, 2012, 12(4): 107-133. http://transport.chd.edu.cn/article/id/201204014 LOU Lu, ZHAO Ling, GENG Tao. Detecting and tracking method of movingvehicle[J]. Journal of Traffic and Transportation Engineering, 2012, 12(4): 107-133. (in Chinese). http://transport.chd.edu.cn/article/id/201204014
[5]	ZIVKOVIC Z, HEIJDEN F. Efficient adaptive density estimation per image pixel for the task of background subtraction[J]. Pattern Recognition Letters, 2006, 27(7): 773-780. doi: 10.1016/j.patrec.2005.11.005
[6]	LEE D S. Effective Gaussian mixture learning for video background subtraction[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(5): 827-832. doi: 10.1109/TPAMI.2005.102
[7]	冯华文, 龚声蓉, 刘纯平. 基于改进高斯混合模型的前景检测[J]. 计算机工程, 2011, 37(19): 179-182. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201119061.htm FENG Hua-wen, GONG Sheng-rong, LIU Chun-ping. Foreground detection based on improved Gaussian mixture model[J]. Computer Engineering, 2011, 37(19): 179-182. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201119061.htm
[8]	刘静, 王玲. 混合高斯模型背景法的一种改进算法[J]. 计算机工程与应用, 2010, 46(13): 168-170. https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG201013052.htm LIU Jing, WANG Ling. Improved algorithm of Gaussian mixture model for background subtraction[J]. Computer Engineering and Applications, 2010, 46(13): 168-170. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG201013052.htm
[9]	李红波, 唐培竣, 吴渝. Kalman滤波器对混合高斯背景建模的改进[J]. 计算机工程与应用, 2009, 45(24): 162-164, 245. https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG200924050.htm LI Hong-bo, TANG Pei-jun, WU Yu. Mixture Gaussian background modeling improved by Kalman filtering[J]. Computer Engineering and Applications, 2009, 45(24): 162-164, 245. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG200924050.htm
[10]	陈祖爵, 陈潇君, 何鸿. 基于改进的混合高斯模型的运动目标检测[J]. 中国图象图形学报, 2007, 12(9): 1585-1589. doi: 10.3969/j.issn.1006-8961.2007.09.014 CHEN Zu-jue, CHEN Xiao-jun, HE Hong. Moving object detection based on improved mixture Gaussian models[J]. Journal of Image and Graphics, 2007, 12(9): 1585-1589. (in Chinese). doi: 10.3969/j.issn.1006-8961.2007.09.014
[11]	SEKI M, WADA T, FUJIWARA H, et al. Background subtraction based on cooccurrence of image variations[C]∥IEEE. Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington DC: IEEE, 2003: 65-72.
[12]	HEIKKILA M, PIETIKAINEN M. A texture-based method for modeling the background and detecting moving objects[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(4): 657-662.
[13]	CHEN Y T, CHEN C S, HUANG C R, et al. Efficient hierarchical method for background subtraction[J]. Patten Recognition, 2006, 40(10): 2706-2715.
[14]	常晓夫, 张文生, 董维山. 基于多种类视觉特征的混合高斯背景模型[J]. 中国图象图形学报, 2011, 16(5): 829-834. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201105018.htm CHANG Xiao-fu, ZHANG Wen-sheng, DONG Wei-shan. Mixture of Gaussian background modeling method based on multi-category visual features[J]. Journal of Image and Graphics, 2011, 16(5): 829-834. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201105018.htm
[15]	MASON M, DURIC Z. Using histograms to detect and track objects in color video[C]∥IEEE. Proceedings of the Applied Imagery Pattern Recognition Workshop. New York: IEEE, 2001: 154-159.
[16]	NORIEGA P, BASCLE B, BERNIER O. Local kernel color histograms for background subtraction[C]∥ALPESH R, HELDER A, BRUNO E. Proceedings of the First International Conference on Computer Vision Theory and Applications. Setubal: INSTICC, 2006: 213-219.
[17]	TSAI T, HUANG Y L, CHIANG T. Image retrieval based on dominant texture features[C]∥IEEE. Proceedings of the IEEE International Symposium on Industrial Electronics. New York: IEEE, 2006: 441-446.
[18]	王永忠, 梁彦, 潘泉, 等. 基于自适应混合高斯模型的时空背景建模[J]. 自动化学报, 2009, 35(4): 371-378. https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO200904008.htm WANG Yong-zhong, LIANG Yan, PAN Quan, et al. Spatiotemporal background modeling based on adaptive mixture of Gaussians[J]. Acta Automatica Sinica, 2009, 35(4): 371-378. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-MOTO200904008.htm