在计算机视觉中，对严重扭曲的景象序列进行三维重建十分困难，本文提出一种简单且有效的迭代方法：首先提出一种基于B样条的非刚体图像配准算法，并结合最优图像块选择算法对扭曲图像序列进行复原；然后提出一种新的L2-范数优化准则，在此基础上结合基于Sampson近似的迭代算法实现多视三角化；最后进行实验并与传统方法比较，实验分析表明，本文提出的一种对扭曲景象序列三维重建迭代方法可以更有效地对扭曲景像序列实现三维重建，并具有算法复杂性小、运算时间短、模拟精度高的优点。%In computer vision, it is very dififcult to perform 3D reconstruction on the severely distorted image series. The paper proposes a simple and effective iteration method that is a non-rigid image registration algorithm based on B-spline combining the optimal image block selection algorithm to restore distorted image series. A new L2 - norm optimization criterion is then introduced and it is combined with an iterative algorithm based on Sampson approximation to realize multi-view triangulation. The study shows that the method to perform 3D reconstruction on distorted image series is more effective compared with the traditional method. This method has the advantages of low complexity of the algorithm, short computing time and high simulation precision.
展开▼
