Home > Published Issues > 2024 > Volume 12, No. 1, 2024 >
JOIG 2024 Vol.12(1): 53-65
doi: 10.18178/joig.12.1.53-65

Localization of Defects on PCBs with Diverse Acquisitions

A. Sainath Chaithanya* and L. Nirmala Devi
Department of Electronics and Communication Engineering, University College of Engineering, Osmania University, Hyderabad, Telangana, India
Email: chaitanya.aravalli@gmail.com (A.S.C.); nirmaladevi@osmania.ac.in (L.N.D.)
*Corresponding author

Manuscript received August 24, 2023; revised October 16, 2023; accepted November 16, 2023; published February 15, 2024.

Abstract—The Printed Circuit Board (PCB) accommodates various Integrated Circuit (IC) components arranged in a specific layout of bond pads, lines, and tracks. Throughout the manufacturing process, irregularities or defects often occur during drilling, etching, stripping, and other stages, impacting the performance and functionality of the circuit board. Many of these defects are related to soldering pads and copper balance, identifying them through manual inspection is time-consuming and error-prone. This necessitates the use of Automated Optical Inspection (AOI). Practices like template matching often require two identical PCBs, which are compared using mathematical algorithms to detect differences. However, they are not resilient to viewpoint variations and non-rigid deformations. The current inspection process primarily focuses on rectifying PCB images captured with tilts ranging from 0 to 84 using homography principles. This correction process operates within a maximum run time of 7.96 s. The adjusted images then undergo analysis via a pattern-matching unit, where the system receives images of the same PCB pattern, each exhibiting different defects. Structural information mapping is performed using various spatial-domain feature-based matching algorithms. When evaluated using SSI and MSE metrics, the model achieved high matching percentages of 99.67%, 99.75%, and 99.30%, and low error rates of 0.343%, 0.358%, and 0.721% for three different types of PCB designs considered. Additionally, the model excels in precisely identifying the location of defects in the PCB images without using bounding boxes, in accordance with the description of the co-images through a segmentation approach. Overall, the proposed system effectively corrects skew, accurately detects abnormalities and outperforms traditional assessment systems.

Keywords—contours, homography, image comparison techniques, image registration, image segmentation, Printed Circuit Board (PCB) defects

Cite: A. Sainath Chaithanya and L. Nirmala Devi, "Localization of Defects on PCBs with Diverse Acquisitions," Journal of Image and Graphics, Vol. 12, No. 1, pp. 53-65, 2024.

Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.