计算机应用 ›› 2015, Vol. 35 ›› Issue (1): 220-223.DOI: 10.11772/j.issn.1001-9081.2015.01.0220

• 虚拟现实与数字媒体 • 上一篇    下一篇

基于相对定向和小区域融合的视频图像拼接

杜丙新   

  1. 安阳师范学院 传媒学院, 河南 安阳455000
  • 收稿日期:2014-08-14 修回日期:2014-09-19 出版日期:2015-01-01 发布日期:2015-01-26
  • 通讯作者: 杜丙新
  • 作者简介:杜丙新(1982-),男,河南信阳人,讲师,硕士研究生,主要研究方向:计算机图形图像、机器视觉.
  • 基金资助:

    河南省科技厅科技攻关重点项目(112102210373, 122102310523).

Video image mosaic based on relative orientation and small region fusion

DU Bingxin   

  1. School of Media and Communications, Anyang Normal University, Anyang Henan 455000, China
  • Received:2014-08-14 Revised:2014-09-19 Online:2015-01-01 Published:2015-01-26

摘要:

针对视频图像拼接中拼接面选取不当造成的畸变问题和图像融合阶段视差带来的残影问题,借助半角纠正法选取合适的拼接面,使用小区域融合法解决融合残影问题.首先,利用摄影测量的相对定向法计算图像姿态参数;然后,根据姿态参数对待拼接图像进行半角纠正并旋转到同一中介平面;接着,利用匹配的同名点对齐图像;最后,在图像重叠区域的中间部分选取一条形区域作为过渡区并采用渐入渐出法进行融合.实验结果表明,在双摄像头视频拼接系统中,半角纠正法比传统平面拼接法的畸变要小;同时与渐入渐出融合法相比,小区域融合法能避免由视差引起的大面积残影问题.经分析,半角纠正和小区域融合法能有效解决视频拼接中图像变换的畸变和重叠区域融合残影问题.

关键词: 相对定向, 半角修正, 小区域融合, 视频图像拼接, 视差

Abstract:

Aiming at the problems in video mosaic system that image transformation distortion caused by improper selection of stitching surface and fusion blur caused by image parallax, half angle correction method was proposed to help select proper stitching surface and small region fusion method was applied to solve fusion blur problem. First, image's attitude parameters were calculated by using relative orientation of photogrammetry; then, according to the attitude parameters, images to be stitched were rotated half angle and mapped to the same intermediary plane; next, the images were aligned by matching points; finally, a bar area was selected in the middle of overlapping area as a transition area and the fade fusion method was applied to the area. The experimental results show that, half angle correction method can achieve smaller distortion than the traditional plane stitching method; compared to fade fusion method, small region fusion method can avoid large area of blur problem caused by parallax. So, half angle correction and small region fusion methods can solve image transformation distortion and overlapping area fusion blur problems effectively.

Key words: relative orientation, half angle correction, small region fusion, video image mosaic, parallax

中图分类号: