Abstract
PurposeThe purpose of this paper is to propose a new inspecting algorithm for defect detection on PCB circuits.Design/methodology/approachPCB circuit images were processed by a radon transformation. A Radon histogram was formed and utilized to establish a texture directional characteristic similarity function. Then, a region of the image which contained the same texture directionality feature was segmented. Furthermore, a directionality estimation method is presented. As the circuit was damaged, the directionality was weakened correspondingly. According to principle, the concept of directional intensity was proposed and then used to measure directionality through analysis of the Radon histogram fluctuation. Finally, the defect was detected based on directional intensity.FindingsThe method has been applied to an inspecting system used in practice and it achieved a higher accuracy and efficiency in comparison with similar methods.Research limitations/implicationsAlthough work on highly intensive PCB circuitry inspection and flaw detection is presented, defect classification was not involved although this is also a very important requirement of inspection.Originality/valueThe paper provides a new way to detect PCB circuitry defects based on texture directionality and proposes evaluating the similarity between image texture directionalities using a radon transformation to search the inspected area. As the inspected region was located, the concept of directional intensity was defined to measure texture directionality to identify defects. The new algorithm performs stably and efficiently and is fit for practical application.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call