Printed Circuit Board (PCB) Defect Detection Using Deep Learning Techniques
Author : Dr. Krishnaveni Sakkarapani and Thejashree P
Abstract :
Printed Circuit Boards (PCBs) are crucial components in modern electronic devices. Detecting defects in PCBs is essential for ensuring product reliability. Traditional inspection methods are time-consuming and prone to human error. This paper presents an automated PCB defect detection system using deep learning, specifically YOLOv8, to identify defects such as missing holes, mouse bites, spurs, spurious copper, short circuits, and open circuits. The system operates by processing input PCB images through the YOLOv8 model, which detects and classifies defects with bounding box annotations. Post-detection, a structured defect report is automatically generated, providing detailed information such as defect type, exact location on the PCB, and severity level. A dataset of PCB images is utilized for training the model, followed by performance evaluation using accuracy. The results demonstrate a high detection rate, enabling efficient defect identification and improving quality control in manufacturing.
Keywords :
PCB defect detection, deep learning, YOLOv8, automated inspection, quality control, object detection, manufacturing, defect classification.
