GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery
The extraction of buildings from remote sensing images is a challenging task. However, existing methods are insufficiently accurate because of the diverse types of buildings, large-scale variability, and complex backgrounds in remote sensing images. There are many deficiencies of the extraction resu...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-11-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843222002345 |
| _version_ | 1828100419876814848 |
|---|---|
| author | Dejun Feng Hongyu Chen Yakun Xie Zichen Liu Ziyang Liao Jun Zhu Heng Zhang |
| author_facet | Dejun Feng Hongyu Chen Yakun Xie Zichen Liu Ziyang Liao Jun Zhu Heng Zhang |
| author_sort | Dejun Feng |
| collection | DOAJ |
| description | The extraction of buildings from remote sensing images is a challenging task. However, existing methods are insufficiently accurate because of the diverse types of buildings, large-scale variability, and complex backgrounds in remote sensing images. There are many deficiencies of the extraction results, such as small building omission, building internal discontinuity, blurred boundary, and irregular building appearance extraction. To solve these problems, a global feature capture and cross-layer information interaction network (GCCINet) is proposed, in which the continuous atrous convolution feature enhancement module is designed to capture a larger range of multiscale building feature information by using continuous atrous convolution to generate different sizes of receptive fields, thus alleviating the problem of discontinuities and the overall appearance of irregular buildings. The global high-low feature cross-fusion module reduces the loss of local information to enhance the ability to identify small buildings through the effective cross-fusion of high-low features. The cross-layer refined fusion and boundary refinement module adopts a unique fusion method to form information fusion between different layers, obtain multiscale context information, and further refine boundaries, thus improving the capability of boundary extraction. The WHU Building Dataset and Inria Dataset are selected for validation. The results show that GCCINet performs state-of-the-art (SOTA) compared with other existing methods, and the performance of GCCINet is verified in terms of usability, interference resistance, robustness, and ablation study. Furthermore, a plug-and-play CAFE module is designed, which can introduce a few parameters to other models while improving their performance. |
| first_indexed | 2024-04-11T08:28:23Z |
| format | Article |
| id | doaj.art-df03dcb555c249da9f65f3e69635d4e6 |
| institution | Directory Open Access Journal |
| issn | 1569-8432 |
| language | English |
| last_indexed | 2024-04-11T08:28:23Z |
| publishDate | 2022-11-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj.art-df03dcb555c249da9f65f3e69635d4e62022-12-22T04:34:37ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322022-11-01114103046GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imageryDejun Feng0Hongyu Chen1Yakun Xie2Zichen Liu3Ziyang Liao4Jun Zhu5Heng Zhang6Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, China; State-Province Joint Engineering Laboratory of Spatial Information Technology of High-Speed Rail Safety, Southwest Jiaotong University, Chengdu 610031, Sichuan, ChinaFaculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, ChinaFaculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, China; Corresponding author.Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, ChinaFaculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, ChinaFaculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, ChinaFaculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, China; China Railway Design Corporation, Tianjin 300308, China; National Engineering Research Center for Digital Construction and Evaluation of Urban Rail Transit, Tianjin 300308, ChinaThe extraction of buildings from remote sensing images is a challenging task. However, existing methods are insufficiently accurate because of the diverse types of buildings, large-scale variability, and complex backgrounds in remote sensing images. There are many deficiencies of the extraction results, such as small building omission, building internal discontinuity, blurred boundary, and irregular building appearance extraction. To solve these problems, a global feature capture and cross-layer information interaction network (GCCINet) is proposed, in which the continuous atrous convolution feature enhancement module is designed to capture a larger range of multiscale building feature information by using continuous atrous convolution to generate different sizes of receptive fields, thus alleviating the problem of discontinuities and the overall appearance of irregular buildings. The global high-low feature cross-fusion module reduces the loss of local information to enhance the ability to identify small buildings through the effective cross-fusion of high-low features. The cross-layer refined fusion and boundary refinement module adopts a unique fusion method to form information fusion between different layers, obtain multiscale context information, and further refine boundaries, thus improving the capability of boundary extraction. The WHU Building Dataset and Inria Dataset are selected for validation. The results show that GCCINet performs state-of-the-art (SOTA) compared with other existing methods, and the performance of GCCINet is verified in terms of usability, interference resistance, robustness, and ablation study. Furthermore, a plug-and-play CAFE module is designed, which can introduce a few parameters to other models while improving their performance.http://www.sciencedirect.com/science/article/pii/S1569843222002345Building extractionRemote sensing imageryGlobal featureCross-layer interactionDeep learning |
| spellingShingle | Dejun Feng Hongyu Chen Yakun Xie Zichen Liu Ziyang Liao Jun Zhu Heng Zhang GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery International Journal of Applied Earth Observations and Geoinformation Building extraction Remote sensing imagery Global feature Cross-layer interaction Deep learning |
| title | GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery |
| title_full | GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery |
| title_fullStr | GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery |
| title_full_unstemmed | GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery |
| title_short | GCCINet: Global feature capture and cross-layer information interaction network for building extraction from remote sensing imagery |
| title_sort | gccinet global feature capture and cross layer information interaction network for building extraction from remote sensing imagery |
| topic | Building extraction Remote sensing imagery Global feature Cross-layer interaction Deep learning |
| url | http://www.sciencedirect.com/science/article/pii/S1569843222002345 |
| work_keys_str_mv | AT dejunfeng gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery AT hongyuchen gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery AT yakunxie gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery AT zichenliu gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery AT ziyangliao gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery AT junzhu gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery AT hengzhang gccinetglobalfeaturecaptureandcrosslayerinformationinteractionnetworkforbuildingextractionfromremotesensingimagery |