Limitations of current visual inspection
Dependence on Manpower
The current PCB inspection process requires one inspector per
VRS. This creates a burden of high labor costs if many VRS
equipment are used.
Low Accuracy and Speed
Visual inspection has various issues, such as inconsistencies in
inspection accuracy and speed, arising from varying skills of the
operator and different standards being applied.
Why PCB Inspection is Difficult
Difficulty in Automatic Classification of Defects
The extent of testing by conventional machine vision
technology is limited to simply checking for the presence
of PCB defects. Functions such as addressing detection
errors and/or automatic defect classification are not possible.
This is because the shapes and the standards of PCBs vary
by company, making it difficult to respond to the various
defect criteria using current equipment.
Benefits Offered by SuaKIT
Reduction of PCB Detection Error and Potential
for Automatic Classification
Customers can benefit greatly by utilizing SUALAB's deep
learning technology. For instance, the AI network picks up
PCB images and automatically reduces chances of false
positives for defects, and also enables automatic defect class
ification. The deep learning technology also allows for
flexible responses for varying customer standards.