基于机器视觉元件管脚高度检测系统研究

Research on height detection system based on machine vision element

目前芯片管脚的几何尺寸缺陷检测局限于管脚的共面度、平整度、位置度检测,对芯片管脚高度缺陷检测算法研究较少,且未从三维角度进行检测。利用线激光传感器结合精密的机械传动结构研发一套三维激光扫描自动检测系统,并提出相应管脚高度检测算法:将采集的芯片管脚点云数据三维重建;提出将所重建的三维图像投影变换为二维图像并灰度化,利用连通域裁剪算法快速定位各管脚位置信息对其进行裁剪,对裁剪的管脚频域高斯低通滤波,用最近邻插值拟合并反灰度化;以xoy面为基准面用求解函数最大值法求其各管脚的最值之差即获得芯片管脚高度,将各管脚拼接后三维显示,并与此芯片标准管脚高度进行比对,对不合格的给予剔除。

At present,the chip defect detection of the chip pin is limited to the coplanarity,flatness and position detection of the pin.There are few researches on the detection algorithm of chip pin height defects,and no three-dimensional detection has been conducted.A three-dimensional laser scanning automatic detection system was developed by using a line laser sensor combined with a precision mechanical transmission structure,and a corresponding pin height detection algorithm was proposed:the acquired chip pin point cloud data was reconstructed in three dimensions;the reconstructed 3D image is transformed into a 2D image and gray scaled.The connected domain clipping algorithm is used to quickly locate the position information of each pin to crop it.The nearest neighbor interpolation is used to simulate and degray-scale the cut pin frequency domain gaussian low-pass filter.Using the xoyplane as the reference plane,the maximum value of each pin is obtained by the maximum value of the solution function.The height of the chip pin is three-dimensionally displayed after splicing the pins,and is compared with the standard pin height of the chip,and the unqualified is rejected.