计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (13): 186-190.

• 图形图像处理 • 上一篇    下一篇

复杂环境下的快速逆行检测跟踪算法

梁东升1,赵连超2,陈永康3,何盛鸿2   

  1. 1.广州市地下铁道总公司 工程研发中心,广州 510030
    2.华南理工大学 自动化科学与工程学院,广州 510640
    3.广州复旦奥特科技股份有限公司,广州 510095
  • 出版日期:2015-07-01 发布日期:2015-06-30

Rapid detecting and tracking algorithm of retrograde-motion in complex environment

LIANG Dongsheng1, ZHAO Lianchao2, CHEN Yongkang3, HE Shenghong2   

  1. 1.Engineering R & D Center,Guangzhou Metro Corporation, Guangzhou 510030, China
    2.School of Automation Science and Engineering, South China University of Technology, Guangzhou 510640, China
    3.Guangzhou Fudan Auto Science & Technology Inc., Guangzhou 510095, China
  • Online:2015-07-01 Published:2015-06-30

摘要: 提出了一种基于光流法的复杂环境逆行检测方法。计算每一帧视频图像中运动区域的光流矢量,并用于判断图像中是否存在逆行,这就是逆行的速报;扩充速报区域,并连续计算和跟踪新区域的光流矢量,用于进一步确认该事件是否为真正的逆行事件,这就是逆行的精报。为验证该算法的有效性,与传统的光流法检测逆行的效果进行了对比实验,结果表明其在逆行的检测时间和检测准确率上均比传统光流法表现更优。

关键词: 光流, 逆行, 速报, 精报

Abstract: A retrograde-motion detecting method based on optical flow is proposed for retrograde-motion detection in a complex environment. Firstly, the optical flow of the movement area in each frame is calculated, and used to determine whether there may be retrograde-motion, this is called quick reporting. Then the report area is expanded, and the optical flow in the new area is calculated and tracked continuous to confirm whether it is real retrograde-motion, this is called exact reporting. To verify the effectiveness of the method, a comparison experiment with the traditional optical flow method is conducted. The results show that the better performance in terms of detection time and accuracy can be achieved using this method.

Key words: optical flow, retrograde-motion, quick report, accurate report