【作者】 马红霞； 肖志涛； 宗凯； 马啸；

【Author】 MA Hong-xia, XIAO Zhi-tao, ZONG Kai, MA Xiao (School of Information and Communication Engineering, Tianjin Polytechnic University, Tianjin 300160, China)

【机构】 天津工业大学信息与通信工程学院；

【摘要】 三维重建主要包括特征点配准和摄像机标定。本文采用灰度相关法和RANSAC法剔除误配准点求得基本矩阵,以及利用摄像机自标定得到的内参数矩阵求出本质矩阵,最后采用SFM重建算法求得空间点坐标。实验结果表明,配准算法对于轮廓较为明显,景深较大的物体效果更好,并可以计算出物体的三维空间点坐标。更多还原

【Abstract】 Three-dimensional reconstruction contains matching of feature points and camera calibration. In matching feature points, the gray-related and RANSAC are used to remove mismatching points to get the fundamental matrix. Then the self-calibration of the camera is used to get the inner parameter matrix of camera, and the essential matrix is worked out using the inner parameter matrix and fundamental matrix. Finally, the coordinates of space points are gained by SFM reconstruction algorithm. The results show that matching algorithm is more effective for the objects with clear outline and deep field, and SFM can compute the coordinates of space points in the 3D reconstruction.更多还原