计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (33): 176-179.

• 图形、图像、模式识别 • 上一篇    下一篇

半自动全向图与卫星遥感图快速配准

杨明凡,王 炜,张茂军,陈 旺,徐 玮   

  1. 国防科学技术大学 信息系统与管理学院 系统工程系,长沙 410073
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-11-21 发布日期:2011-11-21

Fast semiautomatic registration method for ground-level omni-directional image and satellite image

YANG Mingfan,WANG Wei,ZHANG Maojun,CHEN Wang,XU Wei   

  1. Department of System Engineering,College of Information System and Management,National University of Defense Technology,Changsha 410073,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-11-21 Published:2011-11-21

摘要: 折反射全向图和卫星遥感图配准属于异构传感器图像配准问题,目前快速有效的解决方法比较少,但应用需求又比较多,鉴于此,提出了一种基于等角原理的半自动快速配准算法。通过全向Hough变换方法提取全向图中建筑物顶部轮廓直线,同时手工找出卫星图中建筑物顶部轮廓直线,通过角度关系结合投票方法计算出全向图可能拍摄位置,并用所有可能位置组成可行集,从可行集中选出最满意解。仿真实验与实景实验均表明该方法快速有效。

关键词: 图像配准, 三维重建, 全向图, 卫星遥感图

Abstract: Registration between catadioptric omni-directional image and satellite remote sensing image has long been a topic of research in computer vision.There has been increasing demand of the registration approach over the last decade.Unfortunately,few methods have been presented to effectively solve this problem.This paper proposes a fast semi-automatic registration method by analyzing angle relations.It extracts the edges of buildings top side from the omni-directional image by using omni-directional Hough transform,and manually marks the corresponding edges on the satellite view image.It computes all possibility positions that the omni-directional image can be captured through a voting process.It locates the best position from all the candidates.It has demonstrated on experiments that this method is accurate and efficient for the registration.

Key words: image registration, 3D reconstruction, omni-directional image, satellite image