摘要
采用图像参考比较法进行缺陷识别的PCB自动光学检测系统对取像精度要求严格,利用机器视觉系统,以PCB圆形基准点为基准对系统进行校准。采用Hough变换检测基准点圆心,为了提高检测速度和精度,对输入图像进行了灰度化、直方图均衡化、高斯滤波和边缘检测4步预处理。Matlab仿真结果表明,该方法检测精度达到1个像素,耗时为毫秒级,实时性好。
Automatic optical inspection system using reference image method for PCB defects inspection has strict requirement of accurate image location. This paper uses a vision system and circular fiducial markers on PCB to adjust AOI systen~ Hough transform is used to detect central coordinates of fiducial markers. To improve the speed and precision of Hough transform, a marker image will be processed with 4 steps pre-process, including image gray processing, histogram equalization, Gaussian filter and edge detection. Simulation result by Maflab shows that the precision of marker center detection reach one pixel. And time consumption is millisecond level with real-time performance.
出处
《电子设计工程》
2012年第13期117-120,共4页
Electronic Design Engineering
基金
广东省教育部产学研结合基金项目(2007A090302071)
关键词
自动光学检测
基准点对准
圆检测
HOUGH变换
图像处理
automatic inspection system
fiducial registration
circle detection
Hough transform
image processing
