Automated Quality Inspection of Printed Circuit Board

EasyChair Preprint 8174

Abstract

As technology gets advanced, more components depend on the printed circuit board (PCB), and the usage of the PCB layout increases. The tiniest defects on the board might cause serious system harm. PCB surface inspection and identification of defects are one of the most crucial quality control processes. We're using a new model named YOLO-v5 in this procedure, which is designed to locate and detect a variety of PCB defects. This YOLO-v5 algorithm was chosen because of the model's excellent efficacy, precision, and speed. In this paper, we used data that contain 700 images with 4 different types of defects. With a batch size of 16 and a trained epoch of 200, this model achieved a defect detection accuracy of 95.25 percent in PCB.

Keyphrases: Convolution Neural Network, PCB, YOLO v5, deep learning, printed circuit board

@booklet{EasyChair:8174,
  author    = {S D Ullas and Ravikumar Beeranur},
  title     = {Automated Quality Inspection of Printed Circuit Board},
  howpublished = {EasyChair Preprint 8174},
  year      = {EasyChair, 2022}}