CAJ | 학술논문

基本矩阵估计是计算机视觉的关键问题,它是主动立体视觉系统的基础.提出了一种新的估计基本矩阵的线性迭代算法.该算法基于像点到极线的实际距离,利用两个不同视点处的图像先求得两个基本矩阵,再将这两个基本矩阵整合为一个最终的基本矩阵,该矩阵满足秩2约束.基于真实图像的实验表明该算法提高了基本矩阵的估计精度和鲁棒性.
기본구진고계시계산궤시각적관건문제,타시주동입체시각계통적기출.제출료일충신적고계기본구진적선성질대산법.해산법기우상점도겁선적실제거리,이용량개불동시점처적도상선구득량개기본구진,재장저량개기본구진정합위일개최종적기본구진,해구진만족질2약속.기우진실도상적실험표명해산법제고료기본구진적고계정도화로봉성.
Fundamental matrix estimation is a key problem in computer vision and which is the basis of active vision.A new linear iterative algorithm for estimating the fundamental matrix is presented.The algorithm is based on the actual distance between the image point and the epipolar line.It uses the images from different view points to calculate two fundamental matrixes,and then the final fundamental matrix which sailstied the rank-two constraint is calculated from the two matrixes.The experiments on real images show that the algorithm is more accurate and robust.