摘要
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不高,而且运算量大.用双域特征法与先粗后细、粗细结合的搜索法可使得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.实验结果显示了本算法的鲁棒性、快速性与良好拼接效果.
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 cannot deal with it. The information of time domain and the phase domain of images have been fully used with two region feature methods and search method of coarse-fine two steps. It has reduced calculation while it has raised precision which can satisfy the working requirement of automation line. Experimental results show this approach is effective, robust and quick.
出处
《常熟理工学院学报》
2008年第8期105-109,共5页
Journal of Changshu Institute of Technology
关键词
图像拼接
双域
图像匹配
机器视觉
image mosaic
two domains
image registration
machine vision
