利用三视匹配元进行多视影像批处理重建
发布时间：2018-09-03 16:18
【摘要】：无序多视影像的三维重建对噪声非常敏感,错误的匹配关系会影响重建的精度,甚至直接导致重建失败。提出了一种稳健的批处理重建算法,首先利用回路闭合约束剔除可能存在误匹配的三视匹配元,然后以三视匹配元中的三焦张量约束代替传统算法的核线约束来计算所有影像旋转矩阵和相机中心位置的全局最优解。重建过程中引入高效的并查集算法来提取多视匹配点,并利用迭代线性三角形算法计算空间点的三维坐标。实验结果表明,所提算法在重建效率和计算精度方面都能取得较好的结果。
[Abstract]:The 3D reconstruction of disordered multi-view images is very sensitive to noise, and the mismatch relationship will affect the accuracy of reconstruction and even lead to the failure of reconstruction directly. In this paper, a robust batch reconstruction algorithm is proposed. Firstly, the loop closure constraint is used to eliminate the tri-view matching elements which may have mismatch. Then the three-focal Zhang Liang constraint in the tri-view matching element is used instead of the kernel constraint of the traditional algorithm to calculate the global optimal solution of the rotation matrix and the camera center position of all images. In the reconstruction process, an efficient parallel search set algorithm is introduced to extract the multi-view matching points, and the iterative linear triangle algorithm is used to calculate the three-dimensional coordinates of the spatial points. The experimental results show that the proposed algorithm can achieve good results in terms of reconstruction efficiency and computational accuracy.
【作者单位】： 信息工程大学地理空间信息学院;
【基金】：国家973计划(2012CB720000)~~
【分类号】：TP391.41
，
本文编号：2220506
[Abstract]:The 3D reconstruction of disordered multi-view images is very sensitive to noise, and the mismatch relationship will affect the accuracy of reconstruction and even lead to the failure of reconstruction directly. In this paper, a robust batch reconstruction algorithm is proposed. Firstly, the loop closure constraint is used to eliminate the tri-view matching elements which may have mismatch. Then the three-focal Zhang Liang constraint in the tri-view matching element is used instead of the kernel constraint of the traditional algorithm to calculate the global optimal solution of the rotation matrix and the camera center position of all images. In the reconstruction process, an efficient parallel search set algorithm is introduced to extract the multi-view matching points, and the iterative linear triangle algorithm is used to calculate the three-dimensional coordinates of the spatial points. The experimental results show that the proposed algorithm can achieve good results in terms of reconstruction efficiency and computational accuracy.
【作者单位】： 信息工程大学地理空间信息学院;
【基金】：国家973计划(2012CB720000)~~
【分类号】：TP391.41
，
本文编号：2220506
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2220506.html
