Independent distortion correction algorithm in vision measurement system

There are some projected problems during the distortion correction process in vision measurement system, such as complex calculation process, minute operating or adjusting. Hence, a novel point-to-point distortion correction algorithm based on virtual template is proposed in this paper. Based on cross-ratio invariability, characteristics of lens distortion and bilinear interpolation algorithm, all ideal coordinates of corresponding image points projected from a single virtual plane template are solved. And the distortion modified values are calculated on the basis of shifting between ideal and actual coordinates. Then, narrow-range rotations and translations of template running trace are executed. So a virtual distortion correction plate with dense feature light spots is formed step by step, which supplies sufficient interpolation nodes for distortion modification. Finally, all kinds of lens distortions located in the whole image plane are corrected point by point by adopting Delaunay triangulation (DT) technology. And a distortion modified data table is set up, which contains ideal coordinates and distortion modified values of all pixels. The experimental results demonstrate that the standard deviation of projection distance deviations acquired by the independent correction method is 0.015 mm, and the maximum error is merely 0.068 mm. The independent method could effectively correct lens distortion and has benefit for improving vision measurement accuracy.