东北大学学报(自然科学版) ›› 2005, Vol. 26 ›› Issue (10): 1002-1005.DOI: -

• 论著 • 上一篇    下一篇

汽车智能驾驶系统中运动图像的实时检测与跟踪

杨英;赵广耀;武华栋   

  1. 东北大学机械工程与自动化学院;东北大学机械工程与自动化学院;东北大学机械工程与自动化学院 辽宁沈阳110004
  • 收稿日期:2013-06-24 修回日期:2013-06-24 出版日期:2005-10-15 发布日期:2013-06-24
  • 通讯作者: Yang, Y.
  • 作者简介:-
  • 基金资助:
    国家自然科学基金资助项目(10402008)

Real-time detection and track of moving image in intelligent vehicle driving system

Yang, Ying (1); Zhao, Guang-Yao (1); Wu, Hua-Dong (1)   

  1. (1) School of Mechanical Engineering and Automation, Northeastern University, Shenyang 110004, China
  • Received:2013-06-24 Revised:2013-06-24 Online:2005-10-15 Published:2013-06-24
  • Contact: Yang, Y.
  • About author:-
  • Supported by:
    -

摘要: 采用CCD摄像机获取运动目标图像序列,用基于背景重建的运动变化,对运动目标检测;通过运动分割并求运动梯度计算出物体运动方向;利用目标的短时相似性,完成对目标的跟踪与识别,并进而实现对目标长时间的稳定跟踪,确定目标的运动规律,实现运动图像实时检测与跟踪的目的.基于上述运动目标检测跟踪算法,利用C++编写了实时处理程序,实现了动态单目标、多目标及不同方向运动目标的检测与跟踪.实验证明,本算法可以实现对单运动目标和多运动目标的实时检测与跟踪,具有很高的鲁棒性,效果良好.

关键词: 运动图像, 智能车辆, 动目标, 检测与跟踪, 分割与识别

Abstract: Discusses an algorithm of moving object detection on the basis of algorithmic feasibility and demonstrates how to realize the real-time motional object detection. The algorithm detects the changing or moving images by difference method with the image background rebuilt. Then combining the continual characteristics of a motional figure array timely and contentedly, the algorithm sets up the historical records of motional foreground images. Processing further the images, the motional direction and speed of objects are calculated to realize the marking and segmented track of motional region. In addition, this algorithm is also suitable for the multiple occurring sometimes. It lays a sound foundation for image follow-up processing and pattern recognition in video technology with high robustness.

中图分类号: