Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects
Steel surface defects have a significant impact on the quality and performance of many industrial products and cause huge economic losses. Therefore, it is meaningful to detect steel surface defects in real time. To improve the detection performance of steel surface defects with variable scales and...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-07-01
|
Series: | Electronics |
Subjects: | |
Online Access: | https://www.mdpi.com/2079-9292/12/14/3090 |
_version_ | 1797589524028588032 |
---|---|
author | Li Zhang Zhipeng Fu Huaping Guo Yange Sun Xirui Li Mingliang Xu |
author_facet | Li Zhang Zhipeng Fu Huaping Guo Yange Sun Xirui Li Mingliang Xu |
author_sort | Li Zhang |
collection | DOAJ |
description | Steel surface defects have a significant impact on the quality and performance of many industrial products and cause huge economic losses. Therefore, it is meaningful to detect steel surface defects in real time. To improve the detection performance of steel surface defects with variable scales and complex backgrounds, in this paper, a novel method for detecting steel surface defects through a multiscale local and global feature fusion mechanism is proposed. The proposed method uses a convolution operation with a downsampling mechanism in the convolutional neural network model to obtain rough multiscale feature maps. Then, a context-extraction block (CEB) is proposed to adopt self-attention learning on the feature maps extracted by the convolution operation at each scale to obtain multiscale global context information to make up for the shortcomings of convolutional neural networks (CNNs), thus forming a novel multiscale self-attention mechanism. Afterwards, using the feature pyramid structure, multiscale feature maps are fused to improve multiscale object detection. Finally, the channel and spatial attention module and the WIOU (Wise Intersection over Union) loss function are introduced. The model achieved 78.2% and 71.9% mAP respectively on the NEU-DET and GC10-DET dataset. Compared to algorithms such as Faster RCNN and EDDN, this method is effective in improving the detection performance of steel surface defects. |
first_indexed | 2024-03-11T01:07:41Z |
format | Article |
id | doaj.art-61e911a7f42d43619d15a9182bdea334 |
institution | Directory Open Access Journal |
issn | 2079-9292 |
language | English |
last_indexed | 2024-03-11T01:07:41Z |
publishDate | 2023-07-01 |
publisher | MDPI AG |
record_format | Article |
series | Electronics |
spelling | doaj.art-61e911a7f42d43619d15a9182bdea3342023-11-18T19:05:42ZengMDPI AGElectronics2079-92922023-07-011214309010.3390/electronics12143090Multiscale Local and Global Feature Fusion for the Detection of Steel Surface DefectsLi Zhang0Zhipeng Fu1Huaping Guo2Yange Sun3Xirui Li4Mingliang Xu5School of Computer and Information Technology, Xinyang Normal University, Xinyang 464000, ChinaSchool of Computer and Information Technology, Xinyang Normal University, Xinyang 464000, ChinaSchool of Computer and Information Technology, Xinyang Normal University, Xinyang 464000, ChinaSchool of Computer and Information Technology, Xinyang Normal University, Xinyang 464000, ChinaSchool of Computer and Information Technology, Xinyang Normal University, Xinyang 464000, ChinaSchool of Computer and Artificial Intelligence, Zhengzhou University, Zhengzhou 450001, ChinaSteel surface defects have a significant impact on the quality and performance of many industrial products and cause huge economic losses. Therefore, it is meaningful to detect steel surface defects in real time. To improve the detection performance of steel surface defects with variable scales and complex backgrounds, in this paper, a novel method for detecting steel surface defects through a multiscale local and global feature fusion mechanism is proposed. The proposed method uses a convolution operation with a downsampling mechanism in the convolutional neural network model to obtain rough multiscale feature maps. Then, a context-extraction block (CEB) is proposed to adopt self-attention learning on the feature maps extracted by the convolution operation at each scale to obtain multiscale global context information to make up for the shortcomings of convolutional neural networks (CNNs), thus forming a novel multiscale self-attention mechanism. Afterwards, using the feature pyramid structure, multiscale feature maps are fused to improve multiscale object detection. Finally, the channel and spatial attention module and the WIOU (Wise Intersection over Union) loss function are introduced. The model achieved 78.2% and 71.9% mAP respectively on the NEU-DET and GC10-DET dataset. Compared to algorithms such as Faster RCNN and EDDN, this method is effective in improving the detection performance of steel surface defects.https://www.mdpi.com/2079-9292/12/14/3090self-attentionsurface defect detectionconvolutional neural networkmultiscale feature fusion |
spellingShingle | Li Zhang Zhipeng Fu Huaping Guo Yange Sun Xirui Li Mingliang Xu Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects Electronics self-attention surface defect detection convolutional neural network multiscale feature fusion |
title | Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects |
title_full | Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects |
title_fullStr | Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects |
title_full_unstemmed | Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects |
title_short | Multiscale Local and Global Feature Fusion for the Detection of Steel Surface Defects |
title_sort | multiscale local and global feature fusion for the detection of steel surface defects |
topic | self-attention surface defect detection convolutional neural network multiscale feature fusion |
url | https://www.mdpi.com/2079-9292/12/14/3090 |
work_keys_str_mv | AT lizhang multiscalelocalandglobalfeaturefusionforthedetectionofsteelsurfacedefects AT zhipengfu multiscalelocalandglobalfeaturefusionforthedetectionofsteelsurfacedefects AT huapingguo multiscalelocalandglobalfeaturefusionforthedetectionofsteelsurfacedefects AT yangesun multiscalelocalandglobalfeaturefusionforthedetectionofsteelsurfacedefects AT xiruili multiscalelocalandglobalfeaturefusionforthedetectionofsteelsurfacedefects AT mingliangxu multiscalelocalandglobalfeaturefusionforthedetectionofsteelsurfacedefects |