Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation
Pixel-based semantic segmentation models fail to effectively express geographic objects and their topological relationships. Therefore, in semantic segmentation of remote sensing images, these models fail to avoid salt-and-pepper effects and cannot achieve high accuracy either. To solve these proble...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/11/3848 |
| _version_ | 1827690781992812544 |
|---|---|
| author | Wei Cui Meng Yao Yuanjie Hao Ziwei Wang Xin He Weijie Wu Jie Li Huilin Zhao Cong Xia Jin Wang |
| author_facet | Wei Cui Meng Yao Yuanjie Hao Ziwei Wang Xin He Weijie Wu Jie Li Huilin Zhao Cong Xia Jin Wang |
| author_sort | Wei Cui |
| collection | DOAJ |
| description | Pixel-based semantic segmentation models fail to effectively express geographic objects and their topological relationships. Therefore, in semantic segmentation of remote sensing images, these models fail to avoid salt-and-pepper effects and cannot achieve high accuracy either. To solve these problems, object-based models such as graph neural networks (GNNs) are considered. However, traditional GNNs directly use similarity or spatial correlations between nodes to aggregate nodes’ information, which rely too much on the contextual information of the sample. The contextual information of the sample is often distorted, which results in a reduction in the node classification accuracy. To solve this problem, a knowledge and geo-object-based graph convolutional network (KGGCN) is proposed. The KGGCN uses superpixel blocks as nodes of the graph network and combines prior knowledge with spatial correlations during information aggregation. By incorporating the prior knowledge obtained from all samples of the study area, the receptive field of the node is extended from its sample context to the study area. Thus, the distortion of the sample context is overcome effectively. Experiments demonstrate that our model is improved by 3.7% compared with the baseline model named Cluster GCN and 4.1% compared with U-Net. |
| first_indexed | 2024-03-10T10:45:54Z |
| format | Article |
| id | doaj.art-7797b2ebcf774d8a87515f142f221fd5 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T10:45:54Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-7797b2ebcf774d8a87515f142f221fd52023-11-21T22:35:23ZengMDPI AGSensors1424-82202021-06-012111384810.3390/s21113848Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic SegmentationWei Cui0Meng Yao1Yuanjie Hao2Ziwei Wang3Xin He4Weijie Wu5Jie Li6Huilin Zhao7Cong Xia8Jin Wang9School of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaSchool of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, ChinaPixel-based semantic segmentation models fail to effectively express geographic objects and their topological relationships. Therefore, in semantic segmentation of remote sensing images, these models fail to avoid salt-and-pepper effects and cannot achieve high accuracy either. To solve these problems, object-based models such as graph neural networks (GNNs) are considered. However, traditional GNNs directly use similarity or spatial correlations between nodes to aggregate nodes’ information, which rely too much on the contextual information of the sample. The contextual information of the sample is often distorted, which results in a reduction in the node classification accuracy. To solve this problem, a knowledge and geo-object-based graph convolutional network (KGGCN) is proposed. The KGGCN uses superpixel blocks as nodes of the graph network and combines prior knowledge with spatial correlations during information aggregation. By incorporating the prior knowledge obtained from all samples of the study area, the receptive field of the node is extended from its sample context to the study area. Thus, the distortion of the sample context is overcome effectively. Experiments demonstrate that our model is improved by 3.7% compared with the baseline model named Cluster GCN and 4.1% compared with U-Net.https://www.mdpi.com/1424-8220/21/11/3848remote sensing imagessemantic segmentationgeo-object prior knowledgegraph neural network |
| spellingShingle | Wei Cui Meng Yao Yuanjie Hao Ziwei Wang Xin He Weijie Wu Jie Li Huilin Zhao Cong Xia Jin Wang Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation Sensors remote sensing images semantic segmentation geo-object prior knowledge graph neural network |
| title | Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation |
| title_full | Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation |
| title_fullStr | Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation |
| title_full_unstemmed | Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation |
| title_short | Knowledge and Geo-Object Based Graph Convolutional Network for Remote Sensing Semantic Segmentation |
| title_sort | knowledge and geo object based graph convolutional network for remote sensing semantic segmentation |
| topic | remote sensing images semantic segmentation geo-object prior knowledge graph neural network |
| url | https://www.mdpi.com/1424-8220/21/11/3848 |
| work_keys_str_mv | AT weicui knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT mengyao knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT yuanjiehao knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT ziweiwang knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT xinhe knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT weijiewu knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT jieli knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT huilinzhao knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT congxia knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation AT jinwang knowledgeandgeoobjectbasedgraphconvolutionalnetworkforremotesensingsemanticsegmentation |