The performance of automated optical inspection (AOI) systems is constantly progressing toward the ultimate goal of detecting 100% of real defects with no false alarms. Still, until this goal is reached, there is an inherent trade-off in every detection algorithm between detection ability and false alarms.This paper examines the statistical basis of applying several algorithms simultaneously as a method of increasing detection ability and reducing false alarms. It analyzes the probabilities of missing a defect as a function of the number of algorithms and the individual detection probabilities.The theoretical analysis is followed by a discussion of practical implementation of complementing algorithm types and examples of their detection capabilities for typical defects.The theoretical analysis is followed by a discussion of practical implementation of complementing algorithm types and examples of their detection capabilities for typical defects.
展开▼
