基于动态圆的鱼眼图像校正算法研究

针对鱼眼图像传统经度坐标校正效果不明显的缺点,提出了一种改进的鱼眼图像二维校正算法。通过对每一条畸变的经线,不断求取其所在圆的圆心和半径,重新计算经线上点的坐标,有效地校正桶形畸变。该算法不用进行鱼眼镜头的标定以及参数的估计,也不用寻找三维空间点信息,以及在各种坐标系之间进行转换。经实验表明,改进算法校正后的平均垂直偏离角度为传统校正算法的1/3,同时具有耗时少的优点,满足实时性的要求,有较强的实用价值。

Based on the dynamic circular of the fisheye image correction algorithms

The 2d fisheye image correction algorithm was meant to improve the traditional longitude coordinates correction in some extent, for each warp distortion, continue to strike at the center of the circle and its radius, recalculated by the line coordinate points, effectively correcting barrel distortion. The algorithm need not fisheye lens calibration, parameter estimation and three dimensional space information.Experiments show that the improved algorithm average vertical offset angle correction after correction algorithm for the traditional 1/3, while having the advantage of less time-consuming to meet real-time requirements, there is a strong practical value .