一种基于主动视觉的非线性摄像机自标定

针对计算机视觉摄像机镜头畸变对精度的影响问题,提出了一种非线性摄像机标定方法。该方法是将非线性模型视为线性模型和畸变项的叠加,然后在主动视觉下利用线性模型下的正交运动的FOE(当物体或摄像机做纯平移运动时图像对应点连线的交点)点的关系,给出非线性模型参数的约束方程,从而实现非线性模型的摄像机自标定。模拟和真实图像实验均表明,该方法是有效的,具有一定的实用价值。

A Nonlinear Camera Self-calibration Approach Based on Active Vision

For the drawbacks of the traditional calibration approaches in complex operation, a nonlinear camera serf-calibration approach is proposed based on active vision. The basic approach is to regard the nonlinear model as the combination of linear model anddistortion item. Using pole principle and orthogonal relation of pure translation motion of camera or object get a nonlinear constraint equation of intrinsic parameter, and solve all of the nonlinear intrinsic parameters according to DFP of variable metric method, so thenonlinear camera self-calibration is realized. On the basis of theoretical analysis, the algorithm is given. In the end, simulation and real experiments are given, and the results show that this approach is feasible.