计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (13): 158-162.

• 图形图像处理 • 上一篇    下一篇

基于对称差分和背景减的运动目标检测

邱光能,全惠敏   

  1. 湖南大学 电气与信息工程学院,长沙 410082
  • 出版日期:2014-07-01 发布日期:2015-05-12

Moving object detection algorithm based on symmetrical-differencing and background subtraction

QIU Guangneng, QUAN Huimin   

  1. College of Electrical and Information Engineering, Hunan University, Changsha 410082, China
  • Online:2014-07-01 Published:2015-05-12

摘要: 通过对视频运动对象特点的分析,提出一种针对静态场景的运动目标检测算法。该算法采用一种改进的时间平均法初始化背景,在有目标的情况下也能构建出可靠的背景,并融合背景减法和多重对称差分法对背景进行自适应更新。实验结果证明,该算法计算简单,对光线变化具有适应性,能够完整地提取运动目标,改善了运动目标的检测效果,具有一定的鲁棒性。

关键词: 运动目标检测, 自适应背景, 多重对称差分, 光线突变

Abstract: An algorithm for moving object detection with a static background is proposed by analyzing characteristic of movement objects in video. The algorithm uses improved time-averaged method to initialize background, it can obtain reliable background in the condition of having moving objects. During background updating, the paper combines background subtraction method and multi-symmetrical-differencing. Experimental results show that the proposed algorithm is quick, robust, adaptive to illumination changes and it improves the effect of moving object detection.

Key words: moving object detection, adaptive background, multi-symmetrical-differencing, sudden illumination changes