Anomaly detection for X-ray of PCB & IC images

• Main Author: Heng, Daryl Ew-Jynn
• Other Authors: Wen Bihan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; PCB defect detection
• Online Access: https://hdl.handle.net/10356/177102
• Institution: Nanyang Technological University

This project investigates the use of deep learning models for defect detection in printed circuit boards and integrated circuits using YOLOv9. We developed a customized neural network model that take binary mask images and identifies defects in each image. The methodology included converting the data annotations to fit YOLOv9’s format, improving model accuracy by selecting appropriate confidence threshold. Results from experiments indicated an improvement in detection precision, reducing false predictions. The study shows that deep learning techniques can be effectively used to improve printed circuit boards and integrated circuits quality control. This work has broad implications for automated manufacturing processes, highlighting the potential for deep learning to improve industrial quality assurance practices.