Authors
Qunpo Liu1, *, Mengke Wang1, Zonghui Liu1, Bo Su1, Naohiko Hanajima2
1School of Electrical Engineering and Automation, Henan Polytechnic University,
2001 Century Avenue, Jiaozuo, Henan 454003, P. R. China
2College of Information and Systems, Muroran Institute of Technology, 27-1
Mizumoto-cho, Hokkaido, Muroran-shi, Hokkaido 050-8585, Japan
*Corresponding author. Email: [email protected]
Corresponding Author
Qunpo Liu
Received 15 October 2020, Accepted 28 March 2021, Available Online 28 May
2021.
DOI
https://doi.org/10.2991/jrnal.k.210521.005
Keywords
Micro-precision glass insulated terminal; improved Faster R-CNN; missing
block detection
Abstract
Micro-precision Glass Insulated Terminals (referred to as glass terminals)
are the core components used in precision electronic equipment and are
often used for electrical connections between modules. As a glass terminal,
its quality has a great influence on the performance of precision electronic
equipment. Due to the limitations of materials and production processes,
some of the glass terminals produced have defects, such as missing blocks,
pores and cracks. At present, most of the defect detection of glass terminals
is done by manual inspection, and rapid detection easily causes eye fatigue,
so it is difficult to ensure product quality and production efficiency.
The traditional defect detection technology is difficult to effectively
detect the very different defects of the glass terminal. Therefore, this
paper proposes to use deep learning technology to detect missing blocks.
First, preprocess the sample pictures of the missing block defects of the
glass terminal, and then train the improved Faster Region-CNN deep learning
network for defect detection. According to the test results, the accuracy
of the algorithm in detecting missing defects in the glass terminal is
as high as 93.52%.
Copyright
© 2021 The Authors. Published by ALife Robotics Corp. Ltd.
Open Access
This is an open access article distributed under the CC BY-NC 4.0 license
(http://creativecommons.org/licenses/by-nc/4.0/).