利用直线特征的定标图像非线性畸变校正

Line-based nonlinear distortion correction of a calibration image

针对管道弯曲度测量系统中定标图像的特点,提出了一种利用图像中直线特征的非线性畸变校正方法。在畸变得到有效校正的前提下,三维空间中的直线投影到图像平面上应该是直线。先提取畸变图像的边缘,利用边缘线段长度作为判据,剔除野值。把边缘的畸变误差总和作为目标函数,通过最小化目标函数获得最优畸变系数,利用该组系数校正原图像的非线性畸变。实验结果表明,该方法的校正精度较高,适用于管道弯曲度高精度的测量,而且易于实现。

In view of a calibration image in a tube camber measurement system, a nonlinear distortion correction method utilizing linear characteristic is presented. Straight lines in the three dimensional world must project to straight lines in the image plane, if the distortion is compensated effectively. The first step of the method consists of extracting edges from the image by using an edge detection method. And many outliers can be eliminated based on the length of edge segments. The sum of the distortion error of each edge candidate is the objective function. The optimal distortion coefficients can be obtained by minimizing the objective function. Thus the nonlinear distortion of the calibration image can be corrected. Experimental results show that the corrected image meets the need of precision of the tube camber measurement system. And the method can easily be applied.