EUVIMEDEUVIMED entdecken
Crossref

Nondestructive Detection of Capacitors Bonding Defects Combining Active Infrared Thermography and Improved YOLOv8

Yifan Huang, Zixian Song, Bo Sun, Lei Nie, Guanglan Liao, Tielin Shi

IEEE Transactions on Instrumentation and Measurement · 2026 · 75 · 1-13

DOI: 10.1109/tim.2026.3682817 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Instrumentation and Measurement, 2026. Verfasst von Yifan Huang und 5 weiteren Autor:innen. Die Arbeit nennt 44 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Hubei Province Science and Technology Innovation Talent Program, International Science and Technology Joint Research Project of Hubei.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Active frequency-selective surface (FSS) is constructed by loading impedance components onto traditional FSS, and its electromagnetic performance is highly dependent on the bonding quality of capacitor solder bumps. Manual visual inspection cannot meet increasingly stringent testing standards. Herein, we demonstrate a nondestructive method for detecting active FSS bonding defects based on active infrared thermal imaging technology. Steady-state thermal conduction models are established for intact and defective bonding bumps to reveal the temperature-response mechanisms induced by bump fracture, capacitor detachment, misalignment, and excessive solder stacking. Finite element analysis is carried out to determine the optimal thermal excitation temperature of $70~^{\circ }$ C. A Wasserstein generative adversarial network with gradient penalty (WGAN-GP) adversarial network is introduced to generate high-fidelity thermal images for a few defect categories, so as to solve the problem of sample imbalance effectively. A higher resolution feature map as a fine-grained detection layer is embedded in the YOLOv8, and an inner scaling intersection over union (Inner_SIoU) loss with a scaling factor is also introduced to improve the bounding-box regression accuracy under multiscale geometric constraints, thus enhancing the feature fusion capability of the YOLOv8. Our method successfully identifies four types of capacitor solder bump defects and improves the accuracy by 6.05%, compared with the original YOLOv8 model, laying the foundation for the reliability testing of active FSS.

Förderung

  • National Natural Science Foundation of China
  • Hubei Province Science and Technology Innovation Talent Program
  • International Science and Technology Joint Research Project of Hubei
  • Fundamental Research Funds for the Central Universities

Zitieren

APA

Huang, Y., Song, Z., Sun, B., Nie, L., Liao, G., & Shi, T. (2026). Nondestructive Detection of Capacitors Bonding Defects Combining Active Infrared Thermography and Improved YOLOv8. IEEE Transactions on Instrumentation and Measurement, 75, 1-13. https://doi.org/10.1109/tim.2026.3682817

Vancouver

Huang Y, Song Z, Sun B, Nie L, Liao G, Shi T. Nondestructive Detection of Capacitors Bonding Defects Combining Active Infrared Thermography and Improved YOLOv8. IEEE Transactions on Instrumentation and Measurement. 2026;75:1-13. doi:10.1109/tim.2026.3682817

BibTeX
@article{Huang2026,
  author = {Huang, Yifan and Song, Zixian and Sun, Bo and Nie, Lei and Liao, Guanglan and Shi, Tielin},
  title = {Nondestructive Detection of Capacitors Bonding Defects Combining Active Infrared Thermography and Improved YOLOv8},
  journal = {IEEE Transactions on Instrumentation and Measurement},
  year = {2026},
  volume = {75},
  pages = {1--13},
  doi = {10.1109/tim.2026.3682817}
}
Kostenlos starten →