An Empirical Study of the Number of Template Effects on the Efficiency of the Prototype of an Automated Logo Inspection Machine

doi: 10.14456/mijet.2019.12

Abstract

A Product packages have logos printed on them. If these logos were defected, manufacturers may lose customers’ reputation. In addition, the automation increasingly plays an important role in manufacturing process since it can replace human workforce and decrease human errors. Therefore, a prototype of an automated logo inspection machine using pyramidal matching technique was designed and constructed for demonstration and test the efficiency of the proposed logo inspection machine. The original prototype utilized only one image of the package as a template, the experimental results show that the demonstrated machine can distinguish between the complete logos on packages and the defective ones even though the position of the packages were tilted up to ±90 degrees. However, if the package positions were 90 or -90 degrees, the accuracy of the template matching was decreased. Therefore, increasing the number of templates used in the pattern matching algorithm could improve the efficiency of the proposed logo inspection machine by utilizing four templates align in 0, +90, -90, and 180 degree instead of only 0 degree position.