An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms
Abstract The railway track extraction using unmanned aerial vehicle (UAV) aerial images suffers from issues such as low extraction accuracy and high time consumption. In response to these problems, this paper presents a lightweight algorithm DA-DeepLabv3 + based on densely connected and attention me...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-024-84937-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585720876761088 |
|---|---|
| author | Yanbin Weng Jie Yang Changfan Zhang Jing He Cheng Peng Lin Jia Hui Xiang |
| author_facet | Yanbin Weng Jie Yang Changfan Zhang Jing He Cheng Peng Lin Jia Hui Xiang |
| author_sort | Yanbin Weng |
| collection | DOAJ |
| description | Abstract The railway track extraction using unmanned aerial vehicle (UAV) aerial images suffers from issues such as low extraction accuracy and high time consumption. In response to these problems, this paper presents a lightweight algorithm DA-DeepLabv3 + based on densely connected and attention mechanisms. Firstly, the lightweight MobileNetV2 network is employed to replace the Xception feature extraction network, thereby reducing the number of model parameters. Secondly, the receptive field is enlarged by cascading atrous convolutions with different dilation rates in the ASPP (atrous spatial pyramid pooling) module, and other feature maps are concatenated using the multi-scale attention module to enhance the extraction accuracy of the model. Finally, a multi-level upsampling module is designed to enhance the accuracy of boundary contour extraction. Furthermore, a dedicated dataset for railway track segmentation was established to train and evaluate the proposed method. The experimental results indicate that DA-DeepLabv3 + demonstrates significant improvement on the railway track segmentation dataset as well as the DeepGlobe dataset. It achieves mIoU scores of 87.52% and 85.01%, along with accuracy rates of 97.59% and 94.84%, respectively. Compared to classical semantic segmentation networks such as U-Net and DeepLabv3 + , DA-DeepLabv3 + achieves higher extraction accuracy and shorter running time. |
| format | Article |
| id | doaj-art-80c5067e54104a8882a33acc60fdfa0a |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-80c5067e54104a8882a33acc60fdfa0a2025-01-26T12:34:24ZengNature PortfolioScientific Reports2045-23222025-01-0115111610.1038/s41598-024-84937-5An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanismsYanbin Weng0Jie Yang1Changfan Zhang2Jing He3Cheng Peng4Lin Jia5Hui Xiang6School of Computer Science, Hunan University of TechnologySchool of Computer Science, Hunan University of TechnologySchool of Rail Transit, Hunan University of TechnologySchool of Rail Transit, Hunan University of TechnologySchool of Computer Science, Hunan University of TechnologySchool of Rail Transit, Hunan University of TechnologySchool of Computer Science, Hunan University of TechnologyAbstract The railway track extraction using unmanned aerial vehicle (UAV) aerial images suffers from issues such as low extraction accuracy and high time consumption. In response to these problems, this paper presents a lightweight algorithm DA-DeepLabv3 + based on densely connected and attention mechanisms. Firstly, the lightweight MobileNetV2 network is employed to replace the Xception feature extraction network, thereby reducing the number of model parameters. Secondly, the receptive field is enlarged by cascading atrous convolutions with different dilation rates in the ASPP (atrous spatial pyramid pooling) module, and other feature maps are concatenated using the multi-scale attention module to enhance the extraction accuracy of the model. Finally, a multi-level upsampling module is designed to enhance the accuracy of boundary contour extraction. Furthermore, a dedicated dataset for railway track segmentation was established to train and evaluate the proposed method. The experimental results indicate that DA-DeepLabv3 + demonstrates significant improvement on the railway track segmentation dataset as well as the DeepGlobe dataset. It achieves mIoU scores of 87.52% and 85.01%, along with accuracy rates of 97.59% and 94.84%, respectively. Compared to classical semantic segmentation networks such as U-Net and DeepLabv3 + , DA-DeepLabv3 + achieves higher extraction accuracy and shorter running time.https://doi.org/10.1038/s41598-024-84937-5Deep learningSemantic segmentationAttention mechanismRailway extractionUAV aerial imagery |
| spellingShingle | Yanbin Weng Jie Yang Changfan Zhang Jing He Cheng Peng Lin Jia Hui Xiang An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms Scientific Reports Deep learning Semantic segmentation Attention mechanism Railway extraction UAV aerial imagery |
| title | An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms |
| title_full | An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms |
| title_fullStr | An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms |
| title_full_unstemmed | An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms |
| title_short | An improved DeepLabv3 + railway track extraction algorithm based on densely connected and attention mechanisms |
| title_sort | improved deeplabv3 railway track extraction algorithm based on densely connected and attention mechanisms |
| topic | Deep learning Semantic segmentation Attention mechanism Railway extraction UAV aerial imagery |
| url | https://doi.org/10.1038/s41598-024-84937-5 |
| work_keys_str_mv | AT yanbinweng animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT jieyang animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT changfanzhang animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT jinghe animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT chengpeng animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT linjia animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT huixiang animproveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT yanbinweng improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT jieyang improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT changfanzhang improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT jinghe improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT chengpeng improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT linjia improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms AT huixiang improveddeeplabv3railwaytrackextractionalgorithmbasedondenselyconnectedandattentionmechanisms |