Improving Road Semantic Segmentation Using Generative Adversarial Network
Road network extraction from remotely sensed imagery has become a powerful tool for updating geospatial databases, owing to the success of convolutional neural network (CNN) based deep learning semantic segmentation techniques combined with the high-resolution imagery that modern remote sensing prov...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9416669/ |
| _version_ | 1826852311957241856 |
|---|---|
| author | Arnick Abdollahi Biswajeet Pradhan Gaurav Sharma Khairul Nizam Abdul Maulud Abdullah Alamri |
| author_facet | Arnick Abdollahi Biswajeet Pradhan Gaurav Sharma Khairul Nizam Abdul Maulud Abdullah Alamri |
| author_sort | Arnick Abdollahi |
| collection | DOAJ |
| description | Road network extraction from remotely sensed imagery has become a powerful tool for updating geospatial databases, owing to the success of convolutional neural network (CNN) based deep learning semantic segmentation techniques combined with the high-resolution imagery that modern remote sensing provides. However, most CNN approaches cannot obtain high precision segmentation maps with rich details when processing high-resolution remote sensing imagery. In this study, we propose a generative adversarial network (GAN)-based deep learning approach for road segmentation from high-resolution aerial imagery. In the generative part of the presented GAN approach, we use a modified UNet model (MUNet) to obtain a high-resolution segmentation map of the road network. In combination with simple pre-processing comprising edge-preserving filtering, the proposed approach offers a significant improvement in road network segmentation compared with prior approaches. In experiments conducted on the Massachusetts road image dataset, the proposed approach achieves 91.54% precision and 92.92% recall, which correspond to a Mathews correlation coefficient (MCC) of 91.13%, a Mean intersection over union (MIOU) of 87.43% and a F1-score of 92.20%. Comparisons demonstrate that the proposed GAN framework outperforms prior CNN-based approaches and is particularly effective in preserving edge information. |
| first_indexed | 2024-04-10T16:31:52Z |
| format | Article |
| id | doaj.art-efec196f5c9246308ffd9a8a924511ba |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2025-02-16T17:05:58Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-efec196f5c9246308ffd9a8a924511ba2025-01-29T00:00:41ZengIEEEIEEE Access2169-35362021-01-019643816439210.1109/ACCESS.2021.30759519416669Improving Road Semantic Segmentation Using Generative Adversarial NetworkArnick Abdollahi0https://orcid.org/0000-0002-1704-4670Biswajeet Pradhan1https://orcid.org/0000-0001-9863-2054Gaurav Sharma2https://orcid.org/0000-0001-9735-9519Khairul Nizam Abdul Maulud3https://orcid.org/0000-0002-9215-2778Abdullah Alamri4Centre for Advanced Modelling and Geospatial Information Systems (CAMGIS), Faculty of Engineering and IT, School of Information, Systems and Modelling, University of Technology Sydney (UTS), Sydney, NSW, AustraliaCentre for Advanced Modelling and Geospatial Information Systems (CAMGIS), Faculty of Engineering and IT, School of Information, Systems and Modelling, University of Technology Sydney (UTS), Sydney, NSW, AustraliaDepartment of Electrical and Computer Engineering, University of Rochester, Rochester, NY, USAEarth Observation Centre, Institute of Climate Change, Universiti Kebangsaan Malaysia, Bangi, MalaysiaDepartment of Geology and Geophysics, College of Science, King Saud University, Riyadh, Saudi ArabiaRoad network extraction from remotely sensed imagery has become a powerful tool for updating geospatial databases, owing to the success of convolutional neural network (CNN) based deep learning semantic segmentation techniques combined with the high-resolution imagery that modern remote sensing provides. However, most CNN approaches cannot obtain high precision segmentation maps with rich details when processing high-resolution remote sensing imagery. In this study, we propose a generative adversarial network (GAN)-based deep learning approach for road segmentation from high-resolution aerial imagery. In the generative part of the presented GAN approach, we use a modified UNet model (MUNet) to obtain a high-resolution segmentation map of the road network. In combination with simple pre-processing comprising edge-preserving filtering, the proposed approach offers a significant improvement in road network segmentation compared with prior approaches. In experiments conducted on the Massachusetts road image dataset, the proposed approach achieves 91.54% precision and 92.92% recall, which correspond to a Mathews correlation coefficient (MCC) of 91.13%, a Mean intersection over union (MIOU) of 87.43% and a F1-score of 92.20%. Comparisons demonstrate that the proposed GAN framework outperforms prior CNN-based approaches and is particularly effective in preserving edge information.https://ieeexplore.ieee.org/document/9416669/GANroad segmentationremote sensingdeep learningU-Net |
| spellingShingle | Arnick Abdollahi Biswajeet Pradhan Gaurav Sharma Khairul Nizam Abdul Maulud Abdullah Alamri Improving Road Semantic Segmentation Using Generative Adversarial Network IEEE Access GAN road segmentation remote sensing deep learning U-Net |
| title | Improving Road Semantic Segmentation Using Generative Adversarial Network |
| title_full | Improving Road Semantic Segmentation Using Generative Adversarial Network |
| title_fullStr | Improving Road Semantic Segmentation Using Generative Adversarial Network |
| title_full_unstemmed | Improving Road Semantic Segmentation Using Generative Adversarial Network |
| title_short | Improving Road Semantic Segmentation Using Generative Adversarial Network |
| title_sort | improving road semantic segmentation using generative adversarial network |
| topic | GAN road segmentation remote sensing deep learning U-Net |
| url | https://ieeexplore.ieee.org/document/9416669/ |
| work_keys_str_mv | AT arnickabdollahi improvingroadsemanticsegmentationusinggenerativeadversarialnetwork AT biswajeetpradhan improvingroadsemanticsegmentationusinggenerativeadversarialnetwork AT gauravsharma improvingroadsemanticsegmentationusinggenerativeadversarialnetwork AT khairulnizamabdulmaulud improvingroadsemanticsegmentationusinggenerativeadversarialnetwork AT abdullahalamri improvingroadsemanticsegmentationusinggenerativeadversarialnetwork |