基于本质矩阵的摄像机自标定方法

An Essential Matrix-based Camera Self-Calibration Method

本质矩阵描述了在摄像机内参数矩阵已知的条件下的对极几何关系,是归一化图像坐标下的基础矩阵。鉴于本质矩阵具有两相等的非零奇异值,提出了一种基于本质矩阵的自标定方法,该方法首先利用本质矩阵这个特性来构造目标函数,考虑到传统非线性优化算法的诸多不足,最后用粒子群优化算法来求解。实验结果表明,该方法精度较高、鲁棒性较强,是一种简单而有效的自标定方法。

The essential matrix, which is a fundamental matrix described in the normalized image coordinate, represents the epipolar geometry relation under the condition of known camera intrinsic parameters. It is well known, two non-zero singular values of the essential matrix must be equal. Therefore, according to the intrinsic property, an essential matrixbased self-calibration approach is for proposed the first time in this paper. First, the objective function is constructed by the intrinsic property of the essential matrix. Second, the particle swarm optimization is used to solve the objective function considering the drawbacks of traditional optimization algorithms. Analytical results show that the proposed method is not only highly accurate but also robust. Consequently, it is a simple but valid self-calibration method.