JRM Vol.11 No.2 pp. 104-111
doi: 10.20965/jrm.1999.p0104


Automated Visual Inspection for Solder Joints of PCB Based on 3-D Image Analysis

Toshifumi Honda*, Hisae Yamamura*, Mineo Nomoto*, Takanori Ninomiya*, Tomoharu Horii** and Yoshio Miyawaki**

*Production Engineering Research Laboratory, Hitachi Ltd. 292 Yoshida-cho Totsuka-ku, Yokohama, 244-0817 Japan

**Telecommunications Division, Hitachi Ltd. 216 Totsuka-cho, Totsuka-ku, Yokohama, 244-8567, Japan

September 30, 1998
March 5, 1999
April 20, 1999
automatic optical inspection, image processing, three-dimensional scanners

Automated visual solder inspection was studied for different printed circuit boards (PCBs) based on 3-D-image analysis. The optical 3-D scanner accurately detects height of a solder joint from differently focused images of a laser spot. Emitted laser light on the solder joint is detected simultaneously by plural sensors whose focal points are set at different height to realize high-speed height acquisition without secondary-reflection problems caused by the specular surface of the solder joint. An inspection algorithm was designed to recognize a solder joint using detected height and the intensity image so that different shapes of solder fillet do not affect inspection performance. A solder joint is classified based on extracted 3-D features of its fillet shape and 3-D location of the lead from its pad. Inspection parameters are automatically generated with inspection parameter autogeneration using electronic component and PCB design data. Evaluation showed the system gave 100% defect detection and very few false alarms (0.13%) using autogenerated inspection parameters. Results show the technique to be promising in actual production lines for different PCBs.

Cite this article as:
Toshifumi Honda, Hisae Yamamura, Mineo Nomoto, Takanori Ninomiya, Tomoharu Horii, and Yoshio Miyawaki, “Automated Visual Inspection for Solder Joints of PCB Based on 3-D Image Analysis,” J. Robot. Mechatron., Vol.11, No.2, pp. 104-111, 1999.
Data files:

*This site is desgined based on HTML5 and CSS3 for modern browsers, e.g. Chrome, Firefox, Safari, Edge, Opera.

Last updated on Oct. 15, 2021