全景摄像机标定

针对一般全景摄像机标定方法需要特殊的装置、设备和仅适用于满足单视点约束情况下的问题,依据从单视点指向三维点的向量投影到图像平面的过程可以用泰勒级数描述的理论,建立了双曲面全景摄像机标定方法的一般传感器参数模型.标定方法能够补偿镜子焦点与摄像机光心间的配置误差,能够应用于单视点约束没有精确满足的情况下.同时算法只需摄像机从不同位置获取棋盘格平面模板图像可以进行标定.摄像机和平面模板可以自由移动,而不需知道运动的先验知识,所以算法灵活方便,而模型的参数通过非线性优化得到,所以精度较高.

Omni-directional camera calibration

In general, omni-directional camera calibration methods require special scene settings, expensive equipment, and a single viewpoint. To solve these problems, a method using a general sensor parameter model for hyperboloid omni-directional camera calibration was created based on the theory that the projection of a 3D real point onto a pixel of the image plane can be described by a Taylor series expansion. This model can compensate for any misalignment between the focal point of the mirror and the camera's optical center, and can be applied when a single viewpoint has not been met precisely. The proposed method only requires camera motion to be restricted to a planar pattern at a few different locations. Either the camera or the planar pattern can be freely moved without apriori knowledge of the motion, so the method is flexible. Model parameters are obtained by nonlinear optimization, giving it high accuracy.