Advanced metrics for high-precision PCB inspection using 3D x-ray reconstruction
The increasing complexity of modern printed circuit boards (PCBs) demands the adoption of advanced 3D x-ray reconstruction techniques for precise, non-destructive assessment. Current metrological frameworks frequently fall short in addressing pivotal challenges such as the identification of micro-defects, the assurance of solder joint integrity, and the assessment of multi-layer alignment. This research introduces an innovative metrological framework specifically designed for PCB evaluation, incorporating two fundamental metrics aimed at enhancing the fidelity and reliability of reconstructions. The Circuit Trace Fidelity Index (CTFI) effectively measures the geometric accuracy of reconstructed circuit pathways by contrasting them with the empirical ground truth through the deployment of Euclidean distance, thereby promoting the precise discernment of complex configurations. The Layer Alignment Consistency Score (LACS) appraises the accuracy of inter-layer alignment in multi-layered PCBs, utilizing the standard deviation of alignment offsets normalized against an acceptable threshold derived from design tolerances, manufacturing capabilities, or functional testing. These metrics are meticulously crafted to integrate effortlessly with existing reconstruction processes, capitalizing on domain-specific expertise and machine learning methodologies to bolster robustness and applicability. With well-defined mathematical formulations and pragmatic strategies for threshold determination, this framework offers a standardized, quantitative methodology for evaluating reconstruction fidelity. By addressing the particular requirements of PCB inspection, encompassing trace geometry accuracy and inter-layer consistency, this groundbreaking metrology guarantees both structural and functional integrity while optimizing quality control protocols.
Authors
- Patrick Craig (ORCID: https://orcid.org/0000-0002-1047-1987)
- Md Mahfuz Al Hasan (ORCID: https://orcid.org/0000-0002-7788-388X)
- Shajib Ghosh (ORCID: https://orcid.org/0000-0003-3523-8177)
- Nitin Varshney
- Navid Asadizanjani (ORCID: https://orcid.org/0000-0002-7097-4463)
- Antika Roy (ORCID: https://orcid.org/0000-0001-7954-5588)
Institutions
- University of Florida (US)
Publication Details
- Published
- 2025-02-21
- DOI
- https://doi.org/10.1117/12.3058653
- Primary Topic
- Industrial Vision Systems and Defect Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00