Automated PCB Defect Identification System using Machine Learning Techniques

Printed Circuit Boards are mostly seen in all electronic products in daily life. PCBs are easily rejected due to different kinds of defects that slow down the manufacturing process. Defects like spurs, shorts, pinholes, and mouse bites cause issues like leakage of current and open circuits, which affects the performance of PCB. PCB manufacturers must give their utmost priority to delivering defect-free products to remain in the market among various competitors and to meet customer satisfaction. PCB analysts are employed in every industry to ensure that the defects are kept to a minimum or negligible level. But manual inspection of PCB defects is really very challenging, time-consuming, and an unreliable process. There are many defects to detect, which are extremely low tolerance for errors, and significant experience is required to reliably recognize and handle such flawed units. An automated solution for detecting and classifying defects in PCBs is in great demand. Automated defect detectors have significant roles in the quality control of PCB manufacturing sectors due to automated inspection adaptions instead of manual inspections. Defect recognition and individual component position detection are the major concerns in PCB inspection system design. Proper positioning of components improves the operating speed and system accuracy. In this article, a full PCB defect Detector used to automate the detection and classifying defects in printed circuit boards is discussed. It can be implemented in local PCB manufacturing industries. The proposed PCB inspection system functions with maximum accuracy in suitable illumination conditions.