U2Net pipeline crack detection method with improved self-attention

Yuanhao Li; Xin Zuo

doi:10.1117/12.3033198

5 July 2024 U2Net pipeline crack detection method with improved self-attention

Yuanhao Li, Xin Zuo

Proceedings Volume 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024); 1318475 (2024) https://doi.org/10.1117/12.3033198
Event: 3rd International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 2024, Kuala Lumpur, Malaysia

Abstract

To solve the problem that the background of pipeline crack image is complicated and small crack is difficult to detect, an improved U2net pipeline crack detection algorithm based on self-attention is proposed. Firstly, expansion convolution is incorporated into the self-attention module to increase model sensitivity field and reduce image accuracy loss. Secondly, the improved self-attention module is integrated into the U2Net model to extract more global and local crack details. Finally, the detail-specific loss function is used to increase the weight of small cracks, thus effectively enhancing the detection ability of small cracks. The simulation experiment was carried out on the pipeline crack data set, and the experimental results showed that the accuracy of the improved U2Net model reached 71.4%, and the comprehensive evaluation index reached 64.0%. The improved U2Net model could accurately identify fine cracks, and the recognition effect was better than that of the U2net-Full and U2NET-Lite models. It has certain generalization and can be used effectively in the field of crack detection.

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

Citation Download Citation

Yuanhao Li and Xin Zuo "U2Net pipeline crack detection method with improved self-attention", Proc. SPIE 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 1318475 (5 July 2024); https://doi.org/10.1117/12.3033198

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
7 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Image segmentation

Convolution

Data modeling

Education and training

Machine learning

Performance modeling

Deep learning

Show All Keywords

Keywords/Phrases

Search In:

Publication Years