Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips
Steel strip acts as a fundamental material for the steel industry. Surface defects threaten the steel quality and cause substantial economic and reputation losses. Roll marks, always occurring periodically in a large area, are put on the top of the list of the most serious defects by steel mills. Es...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-10-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/21/7264 |
| _version_ | 1797511758111309824 |
|---|---|
| author | Qiwu Luo Weiqiang Jiang Jiaojiao Su Jiaqiu Ai Chunhua Yang |
| author_facet | Qiwu Luo Weiqiang Jiang Jiaojiao Su Jiaqiu Ai Chunhua Yang |
| author_sort | Qiwu Luo |
| collection | DOAJ |
| description | Steel strip acts as a fundamental material for the steel industry. Surface defects threaten the steel quality and cause substantial economic and reputation losses. Roll marks, always occurring periodically in a large area, are put on the top of the list of the most serious defects by steel mills. Essentially, the online roll mark detection is a tiny target inspection task in high-resolution images captured under harsh environment. In this paper, a novel method—namely, Smoothing Complete Feature Pyramid Networks (SCFPN)—is proposed for the above focused task. In particular, the concept of complete intersection over union (CIoU) is applied in feature pyramid networks to obtain faster fitting speed and higher prediction accuracy by suppressing the vanishing gradient in training process. Furthermore, label smoothing is employed to promote the generalization ability of model. In view of lack of public surface image database of steel strips, a raw defect database of hot-rolled steel strip surface, CSU_STEEL, is opened for the first time. Experiments on two public databases (DeepPCB and NEU) and one fresh texture database (CSU_STEEL) indicate that our SCFPN yields more competitive results than several prestigious networks—including Faster R-CNN, SSD, YOLOv3, YOLOv4, FPN, DIN, DDN, and CFPN. |
| first_indexed | 2024-03-10T05:51:37Z |
| format | Article |
| id | doaj.art-12af86a8f1bb42089597c5ce70089c8a |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T05:51:37Z |
| publishDate | 2021-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-12af86a8f1bb42089597c5ce70089c8a2023-11-22T21:39:02ZengMDPI AGSensors1424-82202021-10-012121726410.3390/s21217264Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel StripsQiwu Luo0Weiqiang Jiang1Jiaojiao Su2Jiaqiu Ai3Chunhua Yang4School of Automation, Central South University, Changsha 430006, ChinaSchool of Automation, Central South University, Changsha 430006, ChinaSchool of Automation, Central South University, Changsha 430006, ChinaSchool of Computer Science and Information Engineering, Hefei University of Technology, Hefei 230009, ChinaSchool of Automation, Central South University, Changsha 430006, ChinaSteel strip acts as a fundamental material for the steel industry. Surface defects threaten the steel quality and cause substantial economic and reputation losses. Roll marks, always occurring periodically in a large area, are put on the top of the list of the most serious defects by steel mills. Essentially, the online roll mark detection is a tiny target inspection task in high-resolution images captured under harsh environment. In this paper, a novel method—namely, Smoothing Complete Feature Pyramid Networks (SCFPN)—is proposed for the above focused task. In particular, the concept of complete intersection over union (CIoU) is applied in feature pyramid networks to obtain faster fitting speed and higher prediction accuracy by suppressing the vanishing gradient in training process. Furthermore, label smoothing is employed to promote the generalization ability of model. In view of lack of public surface image database of steel strips, a raw defect database of hot-rolled steel strip surface, CSU_STEEL, is opened for the first time. Experiments on two public databases (DeepPCB and NEU) and one fresh texture database (CSU_STEEL) indicate that our SCFPN yields more competitive results than several prestigious networks—including Faster R-CNN, SSD, YOLOv3, YOLOv4, FPN, DIN, DDN, and CFPN.https://www.mdpi.com/1424-8220/21/21/7264surface defect detectionroll markshot-rolled steelfeature pyramid networks (FPN) |
| spellingShingle | Qiwu Luo Weiqiang Jiang Jiaojiao Su Jiaqiu Ai Chunhua Yang Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips Sensors surface defect detection roll marks hot-rolled steel feature pyramid networks (FPN) |
| title | Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips |
| title_full | Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips |
| title_fullStr | Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips |
| title_full_unstemmed | Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips |
| title_short | Smoothing Complete Feature Pyramid Networks for Roll Mark Detection of Steel Strips |
| title_sort | smoothing complete feature pyramid networks for roll mark detection of steel strips |
| topic | surface defect detection roll marks hot-rolled steel feature pyramid networks (FPN) |
| url | https://www.mdpi.com/1424-8220/21/21/7264 |
| work_keys_str_mv | AT qiwuluo smoothingcompletefeaturepyramidnetworksforrollmarkdetectionofsteelstrips AT weiqiangjiang smoothingcompletefeaturepyramidnetworksforrollmarkdetectionofsteelstrips AT jiaojiaosu smoothingcompletefeaturepyramidnetworksforrollmarkdetectionofsteelstrips AT jiaqiuai smoothingcompletefeaturepyramidnetworksforrollmarkdetectionofsteelstrips AT chunhuayang smoothingcompletefeaturepyramidnetworksforrollmarkdetectionofsteelstrips |