Paper Title
Combination of Deep Learning and Machine Vision for Defect Detection of PCB Assembly

This study proposes automated inspection methods via deep learning object detection algorithms and machine vision template matching technology to solve the issue of defects in printed circuit board assembly. In the current printed circuit board assembly process, assembly defects such as missing components, wrong components, and wrong polarity often occur due to increases in component density, yield and production speed. At present, these defects are inspected manually, and inspection results often have errors or misses. Therefore, this study uses automatic optical inspection (AOI) technology to replace the conventional manual visual inspections. This method primarily aims to detect defects in the solder joints and circuits of printed circuit boards. To address the issue of defects in solid capacitor assembly on printed circuit boards, this study proposes solutions that involve using the YOLO (You Only Look Once) object detection algorithm to find the position of solid capacitors on printed circuit boards, identify their color and quantity, and scan images captured by machine vision template matching technology to locate defects. Keywords - Deep Learning, Machine Vision, Printed Circuit Boards, Electronic Components, You Only Look Once.