Defect Detection Of Printed Circuit Board Assembly Based On Yolov5

This article discusses the development of a new method, PCBA-YOLO, for detecting defects in printed circuit board assemblies (PCBAs). The method utilizes a YOLOv5-based approach with modifications to improve accuracy and efficiency. A new dataset, PCBA-DET, is also introduced for training and testing the model. Results show that the improved method outperforms previous detection networks in terms of accuracy and speed.