摘要
在采用机器视觉技术的挠性印制电路缺陷检测工作中,对基准点的准确定位是实现高质量检测的先决条件。本文选取焊盘圆心或中心作为基准点,针对焊盘的局部形变对基准点定位产生的不良影响,提出了量化目标轮廓边缘方向性的表达方法——方向熵,基于该方法能够获取轮廓坐标与方向参数;接着,根据焊盘模板建立焊盘匹配评估函数筛选出正确的、有价值的、局部的轮廓,并基于轮廓特征实现对焊盘的匹配与对基准点的定位。通过现场测试表明,本文提出的算法能够有效地纠正基准点的定位偏差,将其应用于坐标系变换的目标搜索时,平均定位误差可控制在60um以内,对于后续的目标匹配,缺陷识别与分析具有重大意义。
Location of marker points is an essential previous work in FPC （ Flexible print board） automatic inspection process. The center of the FPC solder is selected as marker point. Aiming to the solder＇ s deformation when locating the marker points, the concept of directional entropy is proposed to describe the distribution of inspecting objective contour＇ s direction, which is applied in extracting solder＇ s contour and directional feature. Then the contour similarity function is established. The valuable contour is reversed and adopted to match solder, and marker＇ s location is also reahzed finally. Experiments show that the marker＇ s locating accuracy is improved, inspecting objective＇ s positioning error is reduced through image registration based on affine transformation,which is within 60 μm. It is significant for practical applications especially the consequent inspecting work such as defect analysis and defect detection.
出处
《激光与红外》
CAS
CSCD
北大核心
2013年第10期1175-1179,共5页
Laser & Infrared
基金
广东省自然科学基金项目(No.S2012010010368)
2011年梅州市产业技术研究与开发资金计划项目资助
关键词
挠性印刷电路板
基准点
方向熵
定位误差
flexible printed
mark point
directional entropy
positioning error
