应用于彩色线扫描机器视觉系统的图像校正方法

Method of Image Correction for Color Line-Scan Machine Vision System

随着机器视觉技术的发展,彩色线扫描系统在工业在线检测中的应用日趋广泛。在彩色线扫描视觉系统中,由于照明光源的照度非均匀性、光学镜头的渐晕和电荷耦合器件(CCD)相机的非均匀性响应等因素,导致系统采集图像出现灰度不均匀且颜色失真的情况,为后端图像处理带来困难,以致对系统稳定性造成影响。基于上述问题,提出了具有针对性的彩色线扫描系统图像校正方法,使该方法与多项式拟合方法对均匀目标进行对比实验,获取的校正后灰度均值与方差等统计结果更优,更接近目标真实灰度,并已将其应用于印刷电路板(PCB)外观缺陷自动光学检测系统,取得了较好的实际应用效果。

As the development of machine vision technology, the color line-scan system is widely applied in the on-line inspection. Due to the non-uniform distribution of optical intensity, the vignetting of optical lens and the different responsivities of charge coupled device （CCD） camera, non-uniform and color distorted image, which induces the problem of image processing and the stability system, could be output from the color line-scan system. A method used in image correction for color line-scan machine vision system is presented. Compared with the polynomial fittings, the mean and variance result of uniform object obtained by this method is better and closer to the real gray of object. The method has been used in the automatic optical inspection of prited circuit board （PCB）.