计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (6): 194-196.DOI: 10.3778/j.issn.1002-8331.2010.06.057

• 图形、图像、模式识别 • 上一篇    下一篇

一种鲁棒的背景配准方法

聂 烜1,陈怀民2   

  1. 1.西北工业大学 软件与微电子学院,西安 710072
    2.西北工业大学 无人机特种技术国家重点实验室,西安 710065
  • 收稿日期:2008-11-26 修回日期:2009-02-25 出版日期:2010-02-21 发布日期:2010-02-21
  • 通讯作者: 聂 烜

Robust background-registration methodology

NIE Xuan1,CHEN Huai-min2   

  1. 1.College of Software and Microelectronics,Northwestern Polytechnical University,Xi’an 710072,China
    2.National Laboratory of UAV Special Technology,Northwestern Polytechnical University,Xi’an 710065,China
  • Received:2008-11-26 Revised:2009-02-25 Online:2010-02-21 Published:2010-02-21
  • Contact: NIE Xuan

摘要: 针对传统背景配准算法的精度完全依赖于匹配点的可靠性而导致效果不稳定的问题,提出了一种鲁棒的背景配准算法。根据统计意义下多数表决的原理,建立了利用更多点匹配关系的背景配准模型,并用广义逆变换将其转化为满秩的线性问题来求解运动模型参数,得到了较稳定的背景配准结果。

关键词: 图像序列, 全局运动补偿, 背景匹配, 仿射参数模型

Abstract: The efficiency and accuracy of conventional background-registration methods usually depend on reliability of matching points,leading to the problem of instability.A robust background-matching algorithm is proposed.Based on the principle of statistical majority,this algorithm establishes a new background-matching model that utilizes more points to implement the matching operation.Furthermore,generalized inverse transform theory is used to the model to full-rank linear problem so as to solve the parameters of mobility scenario and more stable outcome is obtained.

Key words: image sequence, global mobility compensation, background-registration, affine parameter model

中图分类号: