基于网格图像自动检测的镜头畸变校正

Lens Distortion Correction Based on Grid Image Automatic Detection

针对广角镜头所采集图像存在的畸变问题,提出了一种基于模板匹配和质心检测的高精度控制点检测法,并采用基于除法模型和正交相干相位检测的两步校正算法对畸变网格图像进行了高精度校正。实验结果表明,该镜头畸变校正算法能克服光照等不定因素,对畸变的网格图像交叉控制点进行高精度自动提取,并自动计算畸变图像中每一点的校正参数,实现了镜头畸变的高精度、全自动校正。

In order to correct a distorted image collected by the wide-angle lens, a method of precision control point detection based on pattern match and centroid detection is proposed, and using two-step correction based on least-squares fitting and orthogonal coherent phase detection to calibrate the distortion grid image with high precision. Experimental resuhs show thai the proposed method overcome the light and other uncertainties, be valid for high-precision automatic extraction of crossover distortion control point grid image, and automatic calculate correction parameters of each point in the distorted image, realize the high precision and fully automated of the lens distortion correction.