基于深度图像的SMT焊膏印刷实时自动检测

详细信息

Automatic Inspection for SMT Solder Paste Deposition Based on Deep Image

摘要: 表面贴片安装技术(SMT)焊膏印刷质量机器视觉检测中，对采集的图像直接分割精度不高，而印制电路板(PCB)的焊盘、焊膏印刷模板中包含了焊膏的位置和高度信息，利用其建立标准模板节省了统计建模时间。通过焊膏印刷前裸印制线路板(PWB)的数字图像差运算来获取焊膏分割图像，大大提高了检测精度和速度。图像的形态学处理可以消除图像间的正常偏差噪声。图像深度值的提取是利用和无激光的背景图像的差运算。对比实验证明了本方案的检测准确度和速度都有较大提高。
- 机器视觉 /
- 深度图像 /
- 自动检测 /
- 表面贴片安装技术 /
- 焊膏印刷
Abstract: In surface mounted technology(SMT) solder paste deposition automatic vision inspection, statistical pattern establishment takes long time and direct image segmentation suffers bad accuracy. However, the required position and height information can be extracted from the pads on the printed circuit board(PCB) and the stencil which is for solder paste deposition. The segmented images can be obtained by contrasting to the bared PCB images. Morphological processing can eliminate the tolerance between two same PCBs. Comparisons on experiment results showed this approach efficiently improved inspecting accuracy and speed.
- Machine vision /
- Depth image /
- Automatic inspection /
- SMT /
- Solder paste deposition

[1]	周德俭，吴兆华.表面组装工艺技术[M].北京：国防工业出版社,2002.
[2]	罗兵，章云.SMT焊膏印刷质量自动光学检测[J].电子质量,2005,(12):30-32.
[3]	Lathrop R R. Solder paste print qualification using laser triangulation[J]. IEEE Trans on Components, Packaging and Manufacturing Technology,1997,(20):174-182.
[4]	Prasad R P.表面安装技术原理和实践[M].北京：科学出版社,1994.
[5]	Gonzalez R C, Woods R E.数字图像处理[M].北京：电子工业出版社,2003.
[6]	张萍.激光三角法测量关键技术的研究[D].哈尔滨：哈尔滨工业大学,2005.
[7]	刘恩国.基于深度图像的三维视觉关键问题的研究[D].上海：上海交通大学,1999.
[8]	罗兵，曾歆懿，季秀霞.SMT产品缺陷及相应的AOI检测方法[J].电子质量,2006,(12):30-32.