PCB Fault Detection System Using Artificial Intelligence
Abstract
Printed Circuit Boards (PCBs) are essential components of electronic systems, and the identification of manufacturing and assembly defects is important for maintaining product quality and reliability. Conventional manual inspection is time-consuming and can be affected by human error when many boards must be inspected repeatedly. This paper presents a PCB fault detection system using artificial intelligence that integrates a Raspberry Pi 5, an 8 MP USB webcam, OpenCV, and an AI-based detection model for automated inspection. The captured PCB image is preprocessed with OpenCV and analysed using YOLOv8 and based models to identify defects such as broken tracks, scratches, solder bridges, missing or misplaced components, component misalignment, short circuits, surface damage, burn marks, and corrosion. Defective regions are highlighted, and a PASS/FAIL decision is presented through a web dashboard, a 16×2 I²C LCD, green and red LEDs, and a buzzer, while inspection data and images are stored in a database for traceability. A custom interface PCB integrates power distribution, the controller, and indicator circuits, and the prototype was verified through design-rule checking, continuity, power, interface, and integrated system tests. The system is intended to reduce manual inspection effort and support consistent PCB quality monitoring in assembly lines, quality-control laboratories, and consumer, automotive, and medical electronics.