基于双域特征的PCB图像拼接

Printed circuit board image mosaic based on double-domain features

在PCB贴片安装的机器视觉检测中,需要通过图像匹配拼接来得到整个PCB的图像.传统的图像匹配拼接技术精度效果不好,而且运算量大,采用双域特征法和先粗后细、粗细结合的搜索方法可使图像的频域和时域信息得以充分利用,并且在提高精度的同时减少了运算量,能满足流水线工作要求.实验结果显示了该算法的鲁棒性和快速、良好的拼接效果.

In PCB SMT assembly automatic machine vision inspection,image mosaic is necessary for getting an entire PCB image.Because of ineffective precision and large calculation in this engineering image mosaic,conventional mosaic method can t deal with it.The information of time domain and the phase domain of images have used fully with two region feature method and search method of coarse-fine two steps.It has reduced calculation while raised precision which can satisfy the working requirement of automation line.E...