基于变量含误差模型估计基础矩阵

Estimation of the Fundamental Matrix Based on Errors-in-Variables Model

估计基础矩阵是计算机视觉中的重要研究问题。本文提出了一种基于变量含误差(EV)模型的非线性估计方法。建立EV模型之后,本文采用非线性目标函数,并同时估计模型参数与测量误差。此外本文方法还考虑了规范化图像坐标和基础矩阵秩为2的约束。模拟数据和真实图像的实验结果表明,本文方法显著提高了估计基础矩阵的精度。

Estimation of the fundamental matrix is a key problem in computer vision. This paper proposed a nonlinear method to esti-mate it based on Errors-in-Variables (EV) model. After setting up the EV model,this paper adopted nonlinear object function,and estimated both parameters in the model and measurement errors. Besides,this method also considered normalized coordinates and the rank 2 constraint of the fundamental matrix. Experiments of stimulated data and real images show that this method significantly in-cr...