Building an Automatic Defect Verification System Using Deep Neural Network for PCB Defect Classification

Abstract

In the PCB industry, automatic optical inspection (AOI) system takes an important role to increase yield rate. However, the false alarm rate of AOI equipment is high. Therefore, the high cost of human visual inspection at verify and repair system (VRS) station is becoming a problem. Therefore, we propose an automatic defect verification system, called Auto-VRS, to decrease the false alarm rate and reduce operator's workload. The proposed system is composed of two subsystems, referred to fast circuit comparison and deep neural network based defect classification. The fast circuit comparison is to find the accurate defect region of interest (ROI). The deep neural network based defect classification is to verify which is real defect or pseudo defect. The experiment results showed that the Auto-VRS can recognition defects well and has the significant reduction in both false alarm rate and escape rate. With the advantage of the Auto-VRS, it can further improve the VRS operator's efficiency and accuracy in the future.