[1]张嘉烽,张正华,徐 杰,等.基于 Vibe 算法和连通域的车流量统计研究[J].计算机技术与发展,2019,29(12):194-197.[doi:10. 3969 / j. issn. 1673-629X. 2019. 12. 035]
　ZHANG Jia-feng,ZHANG Zheng-hua,XU Jie,et al.Research on Traffic Flow Statistics Based on Vibe Algorithm and Connected Domain[J].,2019,29(12):194-197.[doi:10. 3969 / j. issn. 1673-629X. 2019. 12. 035]
点击复制

基于 Vibe 算法和连通域的车流量统计研究()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
29
期数:
2019年12期
页码:
194-197
栏目:
应用开发研究
出版日期:
2019-12-10

文章信息/Info

Title:
Research on Traffic Flow Statistics Based on Vibe Algorithm and Connected Domain
文章编号:
1673-629X(2019)12-0194-04
作者:
张嘉烽1 ; 张正华1 ; 徐 杰1 ; 房崇鑫 ; 钱 锦1 ; 苏 权2 ; 王 飞2
1. 扬州大学 信息工程学院,江苏 扬州 225127; 2. 扬州国脉通信发展有限责任公司,江苏 扬州 225007
Author(s):
ZHANG Jia-feng 1 ; ZHANG Zheng-hua 1 ; XU Jie 1 ; FANG Chong-xin 1 ; QIAN Jin 1 ; SU Quan 2 ; WANG Fei 2
1. School of Information Engineering,Yangzhou University,Yangzhou 225127,China; 2. Yangzhou Guomai Communication Development Co. ,Ltd. ,Yangzhou 225007,China
关键词:
Vibe 算法; 连通域; 虚拟检测带; 车流统计
Keywords:
Vibe algorithm; connected domain; virtual detection belt; traffic flow statistics
分类号:
TP919.81
DOI:
10. 3969 / j. issn. 1673-629X. 2019. 12. 035
摘要:
随着智慧交通技术的迅速发展,智能车流量统计系统已经成为其中的一个必不可少的环节。 在传统车流量统计方法的基础上,文中改进了一种基于 Vibe 算法和连通域的车流量统计方法,主要是针对车流量统计过程中的背景前景的判断和连通域的扫描进行了研究。 在传统 Vibe 算法的基础上,通过不断实验找到最合适的样本数值和像素半径,从而判断前景车辆并建立背景模型,降低了环境因素的影响。 在选用 8 邻接连通域方法的基础上,逐行扫描像素点并将不邻接的点直接视为背景,优化了基于行程的标记方法,提高了工作效率。 最后在传统虚拟带检测法的基础上,利用 8 行像素作为虚拟带,通过前景的更迭统计出车流量,提高了检测的准确性。 实验结果表明,所用方法的准确性和鲁棒性都较高。
Abstract:
With the rapid development of smart transportation technology,smart traffic statistics system has become an indispensable part. Based on the traditional traffic statistics method,a traffic flow statistics method based on Vibe algorithm and connected domain is improved,which is mainly used to judge the background foreground and scan the connected domain in the process of traffic flow statistics. On the basis of the traditional Vibe algorithm,the most suitable sample value and pixel radius are found throuth continuous experiments,so as to judge the foreground vehicle and establish the background model to reduce the impact of environmental factors. Basedon the method of 8 adjacent connected domains,the pixels are scanned line by line and the non-contiguous points are directly regarded as the background. The stroke-based marking method is optimized to improve the working efficiency. Finally,based on the traditional virtual tape detection method,8 rows of pixels are used as virtual bands,and the traffic flow is calculated by the overlap of the foreground,which improves the detection accuracy. The experiment shows that the method proposed is accurate and robust.

相似文献/References:

[1]王鹏 杜卫东 汪立新 张峰 韩伟 吕志刚.基于连通域特征的射线胶片信息识别技术研究[J].计算机技术与发展,2012,(12):119.
　WANG Peng,DU Wei-dong,WANG Li-xin,et al.Ray Film Character Recognition Technology Research Based on Connected Domain Characteristic[J].,2012,(12):119.
[2]朱家新[],郭德全[],兰时勇[][].基于帧间特征和连通域检测的人数统计[J].计算机技术与发展,2016,26(02):101.
　ZHU Jia-xin[],GUO De-quan[],LAN Shi-yong[][]. People Counting Based on Inter-frame Features and Connected Domain Detection[J].,2016,26(12):101.
[3]陈国军,李 胜.基于空间标记的岩心孔隙并行分割[J].计算机技术与发展,2020,30(12):142.[doi:10. 3969 / j. issn. 1673-629X. 2020. 12. 025]
　CHEN Guo-jun,LI Sheng.Parallel Segmentation of Core Pores Based on Spatial Markers[J].,2020,30(12):142.[doi:10. 3969 / j. issn. 1673-629X. 2020. 12. 025]

更新日期/Last Update: 2019-12-10