【摘要】 提出一种基于未知世界坐标共线特征点的未标定摄像机畸变校正算法以及基于预计算模板的快速实现方法。利用直线的透视投影不变性来估计镜头的畸变参数,不需要知道摄像机的内外部参数,也不需要知道任何空间特征点的世界坐标,简化了镜头畸变校正的过程。该方法的特点是需要的已知条件少,算法简单,易于实施,算法的计算速度快。试验结果表明,该方法能准确快速校正由摄像机镜头畸变引起的图像几何畸变,对机器视觉系统是可行的,预计算模板实现方法大大地降低了计算量,为镜头畸变校正实时应用奠定了基础。更多还原

【Abstract】 A method for distortion correction of un-calibrated lens with unkown collinear feature points is proposed and the improved fast distortion correction calculating method based on pre-computing template is implemented. Without known camera parameters and any known world coordinates point in 3D space, the distortion parameters of camera lens can be estimated based on the perspective projection invariant of straight lines. The proposed distortion correction method is easy to be performed and characterized by simple algorithm, fewer known conditions and fast computing speed. Some experiments to evaluate the performance of this approach on real images are reported, and the results show that this method is feasible for the machine vision systems. The improved method with pre-computing template drastically reduces the computation charge. And real-time applications of the image distortion correction are improved.更多还原