大视场角的鱼眼镜头在成像时由于不同视场垂轴放大率不同, 从而产生畸变使最终成像弯曲变形, 这为识别和测量等应用带来了不便。针对传统经度校正方法的不足, 将算法改进为双向经度鱼眼图像快速校正算法。通过对鱼眼图像有效区域进行划分, 并对不同区域内的畸变点在横、纵两个方向上分别建立校正模型, 确定畸变图像与理想图像之间坐标映射关系, 求取校正坐标的位置。最后对图像进行非线性拉伸, 改善图像中心与边缘放大率不同而产生的“膨胀感”, 获得符合人眼视觉习惯的图像。运用 Matlab对 3组图像进行校正实验, 结果表明, 该算法能快速有效地对鱼眼图像进行畸变校正。

Images taken with large field of view (FOV) fish-eye lenses exhibit distortion because of differing lateral magnification in different fields of view. This kind of distortion makes identification and measurement inconvenient. To resolve the shortcomings of the traditional longitude correction method, an improved rapid bidirectional longitude correction method is proposed. The effective area of the fish-eye image was divided into parts, and a correction model for different points in different vertical and horizontal areas was built to determine the coordinate mapping relationship between the distorted image and the ideal image. The position of the correction coordinates can then be obtained according to this relationship. Finally, nonlinear stretching was performed to abate the “swelling” caused by the difference of magnification between the center and edge of the image to obtain the image that accords with the human-perceived version. Three groups of images were corrected by using MATLAB. The results demonstrate that this method can correct the distortion of fish-eye images quickly and effectively.