摘要： 为快速、高效地校正具有径向畸变的鱼眼图像，提出一种基于几何性质的校正算法。根据投影不变性原理以及径向畸变的几何特性，计算畸变直线的斜率，并通过求解线性方程组得出多项式校正模型的参数。实验结果表明，该算法能够以较低的运算复杂度获得较高的校正精度，相比于采用数学迭代拟合直线的方法，该算法在图像整体校正质量上有明显改善。
关键词:
径向畸变,
几何性质,
畸变直线,
斜率计算,
缩放因子,
图像校正
Abstract: In order to efficiently correct the fisheye image with radial distortion, an algorithm based on geometric properties is proposed. This paper calculates the slope of distortion lines by adopting projection invariance principle and geometric properties of radial distortion. The parameters of polynomial correction model are got by solving linear equations. Experimental result shows this algorithm can get high precision correction with lower complexity. Compared with straight line fitting method using iteration, the overall quality of the correct image is significantly improved when using the algorithm.
Key words:
radial distortion,
geometric property,
distortion lines,
slope calculation,
zoom factor,
image correction
中图分类号:
杨晶晶, 陈更生, 尹文波. 一种基于几何性质的鱼眼图像校正算法[J]. 计算机工程, 2012, 38(3): 203-205.
YANG Jing-Jing, CHEN Geng-Sheng, YIN Wen-Bei. Algorithm for Fisheye Image Correction Based on Geometric Properties[J]. Computer Engineering, 2012, 38(3): 203-205.