DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images
High-resolution remote-sensing images cover more feature information, including texture, structure, shape, and other geometric details, while the relationships among target features are more complex. These factors make it more complicated for classical convolutional neural networks to obtain ideal r...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/5/1328 |
| _version_ | 1797614423142039552 |
|---|---|
| author | Yang Yang Junwu Dong Yanhui Wang Bibo Yu Zhigang Yang |
| author_facet | Yang Yang Junwu Dong Yanhui Wang Bibo Yu Zhigang Yang |
| author_sort | Yang Yang |
| collection | DOAJ |
| description | High-resolution remote-sensing images cover more feature information, including texture, structure, shape, and other geometric details, while the relationships among target features are more complex. These factors make it more complicated for classical convolutional neural networks to obtain ideal results when performing a feature classification on remote-sensing images. To address this issue, we proposed an attention-based multiscale max-pooling dense network (DMAU-Net), which is based on U-Net for ground object classification. The network is designed with an integrated max-pooling module that incorporates dense connections in the encoder part to enhance the quality of the feature map, and thus improve the feature-extraction capability of the network. Equally, in the decoding, we introduce the Efficient Channel Attention (ECA) module, which can strengthen the effective features and suppress the irrelevant information. To validate the ground object classification performance of the multi-pooling integration network proposed in this paper, we conducted experiments on the Vaihingen and Potsdam datasets provided by the International Society for Photogrammetry and Remote Sensing (ISPRS). We compared DMAU-Net with other mainstream semantic segmentation models. The experimental results show that the DMAU-Net proposed in this paper effectively improves the accuracy of the feature classification of high-resolution remote-sensing images. The feature boundaries obtained by DMAU-Net are clear and regionally complete, enhancing the ability to optimize the edges of features. |
| first_indexed | 2024-03-11T07:12:10Z |
| format | Article |
| id | doaj.art-30bfdd71f17a44e9a374973a6468b46e |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T07:12:10Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-30bfdd71f17a44e9a374973a6468b46e2023-11-17T08:31:40ZengMDPI AGRemote Sensing2072-42922023-02-01155132810.3390/rs15051328DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing ImagesYang Yang0Junwu Dong1Yanhui Wang2Bibo Yu3Zhigang Yang4College of Resource Environment and Tourism, Capital Normal University, Beijing 100048, ChinaCollege of Resource Environment and Tourism, Capital Normal University, Beijing 100048, ChinaCollege of Resource Environment and Tourism, Capital Normal University, Beijing 100048, ChinaCollege of Resource Environment and Tourism, Capital Normal University, Beijing 100048, ChinaSurveying and Mapping Institute, Lands and Resource Department of Guangdong Province, Guangzhou 510670, ChinaHigh-resolution remote-sensing images cover more feature information, including texture, structure, shape, and other geometric details, while the relationships among target features are more complex. These factors make it more complicated for classical convolutional neural networks to obtain ideal results when performing a feature classification on remote-sensing images. To address this issue, we proposed an attention-based multiscale max-pooling dense network (DMAU-Net), which is based on U-Net for ground object classification. The network is designed with an integrated max-pooling module that incorporates dense connections in the encoder part to enhance the quality of the feature map, and thus improve the feature-extraction capability of the network. Equally, in the decoding, we introduce the Efficient Channel Attention (ECA) module, which can strengthen the effective features and suppress the irrelevant information. To validate the ground object classification performance of the multi-pooling integration network proposed in this paper, we conducted experiments on the Vaihingen and Potsdam datasets provided by the International Society for Photogrammetry and Remote Sensing (ISPRS). We compared DMAU-Net with other mainstream semantic segmentation models. The experimental results show that the DMAU-Net proposed in this paper effectively improves the accuracy of the feature classification of high-resolution remote-sensing images. The feature boundaries obtained by DMAU-Net are clear and regionally complete, enhancing the ability to optimize the edges of features.https://www.mdpi.com/2072-4292/15/5/1328high-resolution remote-sensing imagesground object classificationdense connectionsmultiscale maximum poolingsemantic segmentation |
| spellingShingle | Yang Yang Junwu Dong Yanhui Wang Bibo Yu Zhigang Yang DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images Remote Sensing high-resolution remote-sensing images ground object classification dense connections multiscale maximum pooling semantic segmentation |
| title | DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images |
| title_full | DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images |
| title_fullStr | DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images |
| title_full_unstemmed | DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images |
| title_short | DMAU-Net: An Attention-Based Multiscale Max-Pooling Dense Network for the Semantic Segmentation in VHR Remote-Sensing Images |
| title_sort | dmau net an attention based multiscale max pooling dense network for the semantic segmentation in vhr remote sensing images |
| topic | high-resolution remote-sensing images ground object classification dense connections multiscale maximum pooling semantic segmentation |
| url | https://www.mdpi.com/2072-4292/15/5/1328 |
| work_keys_str_mv | AT yangyang dmaunetanattentionbasedmultiscalemaxpoolingdensenetworkforthesemanticsegmentationinvhrremotesensingimages AT junwudong dmaunetanattentionbasedmultiscalemaxpoolingdensenetworkforthesemanticsegmentationinvhrremotesensingimages AT yanhuiwang dmaunetanattentionbasedmultiscalemaxpoolingdensenetworkforthesemanticsegmentationinvhrremotesensingimages AT biboyu dmaunetanattentionbasedmultiscalemaxpoolingdensenetworkforthesemanticsegmentationinvhrremotesensingimages AT zhigangyang dmaunetanattentionbasedmultiscalemaxpoolingdensenetworkforthesemanticsegmentationinvhrremotesensingimages |