Abstract A system calibration and part alignment approach for inspection of printed circuitpatterns is presented. We describe procedures for optical and X-Y stage calibration, for axis of rotation estimation, and for correction of nonuniformity in pixel sensitivity. Features are detected with fast and robust techniques that reduce the likelihood of spurious detection. In addition, calibration feature detection is conducted at two different image regions to verify authenticity. This relaxes constraints for a completely contamination-free calibration part. The axis of rotation is calculated by rotating clockwise and counterclockwise and measuring corresponding calibration features. Measurement of rotation during this step and in the alignment procedure is obtained by mapping a quadrilateral in 2-D measurement space onto a rectangle to compensate for minor global variations in part height. This also makes the alignment procedure independent of knowledge of the physical coordinates of the features that define the rectangle's vertices. Analytical solutions for the measurement of rotation and the axis of the rotation are presented. Good results have been obtained in a prototype inspection define the rectangle's vertices. Analytical solutions for the measurement of rotation and the system.
[1]
J. Sanz,et al.
Verifying The Accuracy Of Machine Vision Algorithms And Systems
,
1988,
Twenty-Second Asilomar Conference on Signals, Systems and Computers.
[2]
Anil K. Jain,et al.
Machine-vision techniques for inspection of printed wiring boards and thick-film circuits
,
1986
.
[3]
Theodore D. Doiron.
High precision gaging with computer vision systems
,
1990
.
[4]
Byron Dom,et al.
Machine Vision Techniques for Integrated Circuit Inspection
,
1989
.
[5]
George Wolberg,et al.
Digital image warping
,
1990
.
[6]
John M. Blackwell,et al.
Measurement In Vision - The Need for Calibration
,
1986,
Other Conferences.
[7]
Roland T. Chin,et al.
Automated visual inspection: 1981 to 1987
,
1988,
Computer Vision Graphics and Image Processing.