PCB Defect Detection Using Deep Learning Methods

As a component widely used in electronic products, Printed Circuit Board(PCB) plays an extremely important role in our life. Due to technical limitations, PCB with defects will inevitably appear in the production process. In order to ensure high yield and save labor cost, this paper applied two kinds of target detection network to PCB defect detection and classification tasks. Experiments show that the two methods used in the two different distribution of data sets achieved high accuracy.