首页 | 本学科首页   官方微博 | 高级检索  
     

五点约束最小二乘法估计光流速度场
引用本文:钱东海,张建明.五点约束最小二乘法估计光流速度场[J].计算机辅助设计与图形学学报,2004,16(9):1275-1278.
作者姓名:钱东海  张建明
作者单位:上海大学精密机械工程系,上海,200072
基金项目:上海市教委青年科学基金 ( 0 1QN60 )资助
摘    要:提出一种二维光流场快速计算算法.首先求取当前像素点光流约束线与其8邻域像素点所对应8条光流约束线的交点;其次从8个交点中选取几何位置处于中间的4点,即速度处于中间值,且相互接近的4点,并以其对应的像素点与当前像素点一起构成5置信点;利用5置信点的光流约束方程构造一超定的方程组;最后利用最小二乘法求取当前像素点的光流速度.

关 键 词:计算机视觉  光流场  最小二乘法  像素点  光流约束线  光流速度  图像灰度模式

Computation of Optical Flow by Five-Point-Constraint Least-Squares Method
Qian Donghai,Zhang Jianming.Computation of Optical Flow by Five-Point-Constraint Least-Squares Method[J].Journal of Computer-Aided Design & Computer Graphics,2004,16(9):1275-1278.
Authors:Qian Donghai  Zhang Jianming
Abstract:The algorithm first calculates intersections of optical flow constraint line of the central pixel with constraint lines of its eight-connected pixels. From the eight intersection points, choose four points located in the middle position and close to each other. These four points correspond to four optical flow velocities of middle values. The four chosen points and the central pixel compose five-credible-points and their optical flow constraint equations compose a super determinant equation group. Least-Squares Method is used to calculate the optical flow velocity of the central pixel. such an approach results in higher computing efficiency and robustness.
Keywords:computer vision  optical flow field  least-squares method
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号