Detection of defects on metal surfaces based on PoolFormer Auto-Encoder

Mengyan Li; Guofeng Qin; Jun Zhu; Yongjian Zhu; Peiwen Mi

doi:10.1117/12.3029244

6 May 2024 Detection of defects on metal surfaces based on PoolFormer Auto-Encoder

Mengyan Li, Guofeng Qin, Jun Zhu, Yongjian Zhu, Peiwen Mi

Proceedings Volume 13107, Fourth International Conference on Sensors and Information Technology (ICSI 2024); 131070J (2024) https://doi.org/10.1117/12.3029244
Event: Fourth International Conference on Sensors and Information Technology (ICSI 2024), 2024, Xiamen, China

Abstract

Due to the supervised model data labeling is more complex and cumbersome, this paper adopts unsupervised Auto- Encoder network for defect detection. In addition, for the problem that the convolutional neural network does not have high accuracy in detecting defects on metal surfaces, a defect detection method based on the PoolFormer reconstruction model is proposed. Firstly, according to the type of defects, the defect-free image with superimposed similar morphological noise is used to train the reconstruction model, so that the model has the ability to repair; secondly, the defective image is input into the model for repairing; finally, the residuals between the reconstructed image and the defective image are calculated, and the defect detection and localization can be realized. The experimental results show that the method is able to detect many types of metal surface defects, and the detection accuracy reach more than 98%.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Mengyan Li, Guofeng Qin, Jun Zhu, Yongjian Zhu, and Peiwen Mi "Detection of defects on metal surfaces based on PoolFormer Auto-Encoder", Proc. SPIE 13107, Fourth International Conference on Sensors and Information Technology (ICSI 2024), 131070J (6 May 2024); https://doi.org/10.1117/12.3029244

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
6 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Defect detection

Data modeling

Metals

Machine learning

Transformers

Image restoration

Detection and tracking algorithms

Show All Keywords

Keywords/Phrases

Search In:

Publication Years