摘要
基于挠性印制电路板补强片贴装质量的检测要求,从硬件与软件系统设计出发,开发适用于FPC补强片的贴片质量视觉检测系统。采用自主设计的弧形机架简化标定过程并减少硬件设备对图像采集的影响。在激光三角测量法的基础上,引入高斯拉普拉斯算子和极限式阈值分割用以分离散射光斑的光心和散射光区域,同时采用形态学运算进一步分离出光心区域。运用矩的不变性进行椭圆拟合并将其中心位置的改变转换为补强片厚度的变化,更易于观测。实验结果表明,改变结构元素的大小,光斑中心位置的波动为1.15像素,且最大误差仅达0.82像素,若闭运算结构元素设定为8,能进一步提高检测精度与检测效率。
According to the measuring requirements of mount quality for flexible printed circuit(FPC),starting from the designs of the hardware and software subsystems,an automatic mount quality inspection system based on machine vision was developed.The use of self-designed arc frame was utilized to simplify the process of calibration and reduce the influence of hardware on image acquisition.On the basis of laser triangulation method,the Laplacian of Gaussian operator and ultimate threshold segmentation methods were introduced to distinguish between centre and scattered area for the scattered light spot.And then the centric region was separated out by morphological closing operation.Meanwhile,the invariance of spatial moments was applied to ellipse fitting whose central position was converted to the change of stiffener on thickness that was easier to make observation.The results show that the fluctuation of spot center location is 1.15 pixels.Furthermore,the maximum error of location is only 0.82 pixels.And if the structure element of closed operation is 8,it can further improve the detection accuracy and efficiency.
出处
《应用光学》
CAS
CSCD
北大核心
2015年第6期942-947,共6页
Journal of Applied Optics
基金
国家自然科学基金项目(51205115)
大学生创新创业训练计划项目(201310500007)
关键词
机器视觉
FPC补强片
激光三角测量法
椭圆拟合
machine vision
FPC stiffener
laser triangulation measurement
ellipse fitting