作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2006, Vol. 32 ›› Issue (1): 198-200.

• 人工智能及识别技术 • 上一篇    下一篇

基于梯度光流场计算方法的一种改进

危水根,陈 震,黎 明   

  1. 南昌航空工业学院人工智能和图像处理研究中心, 南昌 330034
  • 出版日期:2006-01-05 发布日期:2006-01-05

An Improvement of Optical Flow Estimation Based on Gradient

WEI Shuigen, CHEN Zhen, LI Ming   

  1. Center of Artificial Intelligence and Image Processing, Nanchang Institute of Aeronautical Technology, Nanchang 330034
  • Online:2006-01-05 Published:2006-01-05

摘要: 提出了一种新的改进光流场技术计算方法。它采用类似差分的方法从图像序列的多帧图像中检测出运动目标区域。通过代数运算减轻了Horn 算法中全局约束对运动边界的影响。实验证明该方法是有效的。

关键词: 图像序列;类似差分法;运动目标区域;光流场

Abstract: An improvement of optical flow estimation based on gradient is presented in this paper. Motion object region is extracted from multi-frame images in image sequences by using similar difference method. The harmful effect on the motion edge by global smoothness constraint in Horn algorithm is weakened by algebraic operation. Experimental results show that this method is suitable and effective.

Key words: Image sequences;Similar difference method;Motion object region;Optical flow field