计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (24): 78-80.

• 学术探讨 • 上一篇    下一篇

一种透视变换图像金字塔匹配改进算法

金勇俊,李言俊,张 科   

  1. 西北工业大学 航天学院,西安 710072
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-08-21 发布日期:2007-08-21
  • 通讯作者: 金勇俊

Improved perspective transformation image pyramid registration method

JIN Yong-jun,LI Yan-jun,ZHANG Ke   

  1. College of Astronautics,Northwestern Polytechnical University,Xi’an 710072,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-08-21 Published:2007-08-21
  • Contact: JIN Yong-jun

摘要: 为提高图像匹配的精度和稳定性,在图像匹配过程中用消除了Hessian矩阵的LMA改进算法对两幅图像透视变换矩阵的8个未知参数进行拟合,减少了迭代的计算量。匹配过程中用多分辨率金字塔法对图像进行分解,采用由粗到精的搜索策略,进一步减少了计算量并避免了误匹配。通过对照片进行匹配测试,证明了这种方法的有效性和实用价值。

关键词: 透视变换, 图像匹配, 图像金字塔, LMA

Abstract: In order to improve the robust and precise in image registration,we get the Hessian matrix removed to improve LMA for searching the eight unknown parameters of two input perspective transformation images and get computational gains.We use multi-resolution pyramid consists of a set of images representing an image in multi-resolution.With the coarsest level to the finest level search strategy,we get large computational gains and help prevent getting trapped in local minima.As the experimental results reveal,this approach is efficient and useful.

Key words: perspective transformation, image registration, pyramid image, Levenberg-Marquardt Aalgorithm(LMA)