A new method for deep learning detection of defects in X-ray images of pressure vessel welds
Abstract Given that defect detection in weld X-ray images is a critical aspect of pressure vessel manufacturing and inspection, accurate differentiation of the type, distribution, number, and area of defects in the images serves as the foundation for judging weld quality, and the segmentation method...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-03-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-024-56794-9 |
| _version_ | 1797259320603181056 |
|---|---|
| author | Xue Wang Feng He Xu Huang |
| author_facet | Xue Wang Feng He Xu Huang |
| author_sort | Xue Wang |
| collection | DOAJ |
| description | Abstract Given that defect detection in weld X-ray images is a critical aspect of pressure vessel manufacturing and inspection, accurate differentiation of the type, distribution, number, and area of defects in the images serves as the foundation for judging weld quality, and the segmentation method of defects in digital X-ray images is the core technology for differentiating defects. Based on the publicly available weld seam dataset GDX-ray, this paper proposes a complete technique for fault segmentation in X-ray pictures of pressure vessel welds. The key works are as follows: (1) To address the problem of a lack of defect samples and imbalanced distribution inside GDX-ray, a DA-DCGAN based on a two-channel attention mechanism is devised to increase sample data. (2) A convolutional block attention mechanism is incorporated into the coding layer to boost the accuracy of small-scale defect identification. The proposed MAU-Net defect semantic segmentation network uses multi-scale even convolution to enhance large-scale features. The proposed method can mask electrostatic interference and non-defect-class parts in the actual weld X-ray images, achieve an average segmentation accuracy of 84.75% for the GDX-ray dataset, segment and accurately rate the valid defects with a correct rating rate of 95%, and thus realize practical value in engineering. |
| first_indexed | 2024-04-24T23:07:33Z |
| format | Article |
| id | doaj.art-ee91e0d331f34b3a9fd09cdc04a9b383 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-24T23:07:33Z |
| publishDate | 2024-03-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-ee91e0d331f34b3a9fd09cdc04a9b3832024-03-17T12:26:24ZengNature PortfolioScientific Reports2045-23222024-03-0114111210.1038/s41598-024-56794-9A new method for deep learning detection of defects in X-ray images of pressure vessel weldsXue Wang0Feng He1Xu Huang2School of Electrical Engineering, Chongqing University of Science and TechnologySchool of Electrical Engineering, Chongqing University of Science and TechnologySchool of Electrical Engineering, Chongqing University of Science and TechnologyAbstract Given that defect detection in weld X-ray images is a critical aspect of pressure vessel manufacturing and inspection, accurate differentiation of the type, distribution, number, and area of defects in the images serves as the foundation for judging weld quality, and the segmentation method of defects in digital X-ray images is the core technology for differentiating defects. Based on the publicly available weld seam dataset GDX-ray, this paper proposes a complete technique for fault segmentation in X-ray pictures of pressure vessel welds. The key works are as follows: (1) To address the problem of a lack of defect samples and imbalanced distribution inside GDX-ray, a DA-DCGAN based on a two-channel attention mechanism is devised to increase sample data. (2) A convolutional block attention mechanism is incorporated into the coding layer to boost the accuracy of small-scale defect identification. The proposed MAU-Net defect semantic segmentation network uses multi-scale even convolution to enhance large-scale features. The proposed method can mask electrostatic interference and non-defect-class parts in the actual weld X-ray images, achieve an average segmentation accuracy of 84.75% for the GDX-ray dataset, segment and accurately rate the valid defects with a correct rating rate of 95%, and thus realize practical value in engineering.https://doi.org/10.1038/s41598-024-56794-9Pressure vessel weldsX-ray imagesDC-GANDefect segmentationU-Net |
| spellingShingle | Xue Wang Feng He Xu Huang A new method for deep learning detection of defects in X-ray images of pressure vessel welds Scientific Reports Pressure vessel welds X-ray images DC-GAN Defect segmentation U-Net |
| title | A new method for deep learning detection of defects in X-ray images of pressure vessel welds |
| title_full | A new method for deep learning detection of defects in X-ray images of pressure vessel welds |
| title_fullStr | A new method for deep learning detection of defects in X-ray images of pressure vessel welds |
| title_full_unstemmed | A new method for deep learning detection of defects in X-ray images of pressure vessel welds |
| title_short | A new method for deep learning detection of defects in X-ray images of pressure vessel welds |
| title_sort | new method for deep learning detection of defects in x ray images of pressure vessel welds |
| topic | Pressure vessel welds X-ray images DC-GAN Defect segmentation U-Net |
| url | https://doi.org/10.1038/s41598-024-56794-9 |
| work_keys_str_mv | AT xuewang anewmethodfordeeplearningdetectionofdefectsinxrayimagesofpressurevesselwelds AT fenghe anewmethodfordeeplearningdetectionofdefectsinxrayimagesofpressurevesselwelds AT xuhuang anewmethodfordeeplearningdetectionofdefectsinxrayimagesofpressurevesselwelds AT xuewang newmethodfordeeplearningdetectionofdefectsinxrayimagesofpressurevesselwelds AT fenghe newmethodfordeeplearningdetectionofdefectsinxrayimagesofpressurevesselwelds AT xuhuang newmethodfordeeplearningdetectionofdefectsinxrayimagesofpressurevesselwelds |