Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention
Cracks inside urban underground comprehensive pipe galleries are small and their characteristics are not obvious. Due to low lighting and large shadow areas, the differentiation between the cracks and background in an image is low. Most current semantic segmentation methods focus on overall segmenta...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-02-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/24/3/1005 |
| _version_ | 1797318166987145216 |
|---|---|
| author | Yuan Zhou Zhiyu Yang Xiaofeng Bai Chengwei Li Shoubin Wang Guili Peng Guodong Li Qinghua Wang Huailei Chang |
| author_facet | Yuan Zhou Zhiyu Yang Xiaofeng Bai Chengwei Li Shoubin Wang Guili Peng Guodong Li Qinghua Wang Huailei Chang |
| author_sort | Yuan Zhou |
| collection | DOAJ |
| description | Cracks inside urban underground comprehensive pipe galleries are small and their characteristics are not obvious. Due to low lighting and large shadow areas, the differentiation between the cracks and background in an image is low. Most current semantic segmentation methods focus on overall segmentation and have a large perceptual range. However, for urban underground comprehensive pipe gallery crack segmentation tasks, it is difficult to pay attention to the detailed features of local edges to obtain accurate segmentation results. A Global Attention Segmentation Network (GA-SegNet) is proposed in this paper. The GA-SegNet is designed to perform semantic segmentation by incorporating global attention mechanisms. In order to perform precise pixel classification in the image, a residual separable convolution attention model is employed in an encoder to extract features at multiple scales. A global attention upsample model (GAM) is utilized in a decoder to enhance the connection between shallow-level features and deep abstract features, which could increase the attention of the network towards small cracks. By employing a balanced loss function, the contribution of crack pixels is increased while reducing the focus on background pixels in the overall loss. This approach aims to improve the segmentation accuracy of cracks. The comparative experimental results with other classic models show that the GA SegNet model proposed in this study has better segmentation performance and multiple evaluation indicators, and has advantages in segmentation accuracy and efficiency. |
| first_indexed | 2024-03-08T03:48:32Z |
| format | Article |
| id | doaj.art-22b010b18bd04b56ad0f6ce9edb5f2c6 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-08T03:48:32Z |
| publishDate | 2024-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-22b010b18bd04b56ad0f6ce9edb5f2c62024-02-09T15:22:33ZengMDPI AGSensors1424-82202024-02-01243100510.3390/s24031005Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global AttentionYuan Zhou0Zhiyu Yang1Xiaofeng Bai2Chengwei Li3Shoubin Wang4Guili Peng5Guodong Li6Qinghua Wang7Huailei Chang8School of Instrument Science and Engineering, Harbin Institute of Technology, Harbin 150001, ChinaSchool of Control and Mechanical, Tianjin Chengjian University, Tianjin 300384, ChinaSchool of Instrument Science and Engineering, Harbin Institute of Technology, Harbin 150001, ChinaSchool of Instrument Science and Engineering, Harbin Institute of Technology, Harbin 150001, ChinaSchool of Control and Mechanical, Tianjin Chengjian University, Tianjin 300384, ChinaSchool of Control and Mechanical, Tianjin Chengjian University, Tianjin 300384, ChinaSTECOL Corporation, Power Construction Corporation of China, Tianjin 300384, ChinaSTECOL Corporation, Power Construction Corporation of China, Tianjin 300384, ChinaSTECOL Corporation, Power Construction Corporation of China, Tianjin 300384, ChinaCracks inside urban underground comprehensive pipe galleries are small and their characteristics are not obvious. Due to low lighting and large shadow areas, the differentiation between the cracks and background in an image is low. Most current semantic segmentation methods focus on overall segmentation and have a large perceptual range. However, for urban underground comprehensive pipe gallery crack segmentation tasks, it is difficult to pay attention to the detailed features of local edges to obtain accurate segmentation results. A Global Attention Segmentation Network (GA-SegNet) is proposed in this paper. The GA-SegNet is designed to perform semantic segmentation by incorporating global attention mechanisms. In order to perform precise pixel classification in the image, a residual separable convolution attention model is employed in an encoder to extract features at multiple scales. A global attention upsample model (GAM) is utilized in a decoder to enhance the connection between shallow-level features and deep abstract features, which could increase the attention of the network towards small cracks. By employing a balanced loss function, the contribution of crack pixels is increased while reducing the focus on background pixels in the overall loss. This approach aims to improve the segmentation accuracy of cracks. The comparative experimental results with other classic models show that the GA SegNet model proposed in this study has better segmentation performance and multiple evaluation indicators, and has advantages in segmentation accuracy and efficiency.https://www.mdpi.com/1424-8220/24/3/1005cracksemantic segmentationattention modelloss functionGA-SegNet |
| spellingShingle | Yuan Zhou Zhiyu Yang Xiaofeng Bai Chengwei Li Shoubin Wang Guili Peng Guodong Li Qinghua Wang Huailei Chang Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention Sensors crack semantic segmentation attention model loss function GA-SegNet |
| title | Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention |
| title_full | Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention |
| title_fullStr | Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention |
| title_full_unstemmed | Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention |
| title_short | Semantic Segmentation of Surface Cracks in Urban Comprehensive Pipe Galleries Based on Global Attention |
| title_sort | semantic segmentation of surface cracks in urban comprehensive pipe galleries based on global attention |
| topic | crack semantic segmentation attention model loss function GA-SegNet |
| url | https://www.mdpi.com/1424-8220/24/3/1005 |
| work_keys_str_mv | AT yuanzhou semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT zhiyuyang semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT xiaofengbai semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT chengweili semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT shoubinwang semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT guilipeng semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT guodongli semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT qinghuawang semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention AT huaileichang semanticsegmentationofsurfacecracksinurbancomprehensivepipegalleriesbasedonglobalattention |