由双目序列影像解求运动参数的非立体匹配方法

基金项目: 国家自然科学基金资助项目

详细信息

A New Approach of 3-D General Motion from Bincular Image Flow without Correspondence

Dept. of Aeropbtrerammetry and Remote Sensing, WTUSM, Luoyu Road, 39, Wuhan, China, 430070

摘要: 在无需实现精确立体匹配的条件下,导出了由双目序列影像得到的双目影像光流场严密解算空间物体运动参数的数学模型,并从理论上分析了该模型的稳健性,同时给出了简单的验证算例。
- 双目序列影像 /
- 光流 /
- 运动参数估计
Abstract: A new approach of 3-D general motion estimation from bincular image flows without correspondence is presented. It is a robust algorthm even if there are some drop-in points and drop-out points in left and right images. A sample experiment illustrate this approach is feasible.
- bincular sequence images /
- optical flow /
- motion parameters estimation