Bare PCB (Printed Circuit Board) Fault Detection in Real-time Using YOLOv5

Abstract

With the continuous development of object detection technology, the YOLO series of algorithms with high precision and speed has been used in various fault detection tasks. This paper aims to establish a real-time quality control framework for PCB fault detection that can be adopted by industries. The development of more automized testing method by using YOLOv5. Detection of imperfections despite misalignment and incorrect orientation is required in the industry. Decreases quality check time and human examiner cost for quality control and auditing the product (as it might be slow and error-prone). Computer Camera is used of real-time fault detection in this prototype, however any digital camera integration would be possible and might improve the performance. Ultimately, increased revenue generation followed by better quality production and supply that meets the market demand.