基于图像处理的PCB在线检测图像校正

On-line Image Correction of PCB Based on Image Processing

印刷电路板(PCB)的在线自动光学检测需要校正因摄像机视场较大等原因产生的PCB光学图像的畸变。提出了一种基于图像处理的PCB在线检测的图像校正方法。首先对畸变校正的基本原理进行了详细介绍,即制作棋盘图标定板,通过拍摄板定板获取获取足够数量不同位置和角度的标定图片,然后对标定图片的分析处理。在张正友法的基础上,利用Open CV自动提取每一张标定图片的角点信息,并进行亚像素提升,计算图像与标定板之间的单应性矩阵,从而求出摄像机的内部参数与外部参数。然后考虑畸变因素,通过最大似然估计得到所有参数,对摄像机进行标定,最后实现对在线PCB光学图像的校正试验。实验结果证明该方法可以较好地实现对在线PCB图像的校正,为进一步分析处理在线PCB图像奠定了基础。

In this paper,an image correction method based on image processing for PCB online detection was proposed.Firstly,the basic principle of distortion correction is introduced in detail.The chessboard is made and a sufficient number of calibration pictures of different positions and angles are obtained.On the basis of Zhang Zhengyou method,the chessboard wsa made and a sufficient number of calibration pictures are obtained.The corner information of each calibration picture was automatically extracted with Open CV,and sub-pixel lifting is carried out.The correction experiment of the online PCB optical image is realized.