The method employs circularly distributed cumulative histograms and neural networks to identify defects in PCB images, suggesting a novel image processing approach.
The presented work considers the approach for detecting defects in printed circuit boards, including flexible ones. It is based on software for image processing and machine learning. The first stage realizes the algorithm of constructing images reflecting circularly distributed cumulative histograms (CDCH) of the standard and sample printed circuit board images. They reflect size-independent statistical features of the input PCB images. The second stage applies the neural network to the CDCH images identified in the first step. The approach allows us to divide the boards into correct and defective without indicating the types of defects, to determine a type of defect as redundant or missing components, or light or dark spots on the board. Neural networks of artificial intelligence, which are available on the Internet, were used in the work.
No takes yet. Share an insight, caveat, or question.
Melnyk et al. (2025) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: