Multi-scale attention integrated hierarchical networks for high-resolution building footprint extraction
文献类型：期刊论文
| 作者 | Liu, Tang2,3; Yao, Ling3,4,5,6; Qin, Jun3; Lu, Ning3,4,5; Jiang, Hou3; Zhang, Fan1; Zhou, Chenghu2,3 |
| 刊名 | INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
 |
| 出版日期 | 2022-05-01 |
| 卷号 | 109页码:11 |
| 关键词 | Building footprint extraction Deep learning Attention mechanism Remote sensing imagery Semantic segmentation |
| ISSN号 | 1569-8432 |
| DOI | 10.1016/j.jag.2022.102768 |
| 通讯作者 | Yao, Ling(yaoling@lreis.ac.cn) |
| 英文摘要 | Convolutional neural networks show excellent performance in image segmentation. However, compared with natural images, remote sensing images are characterized by large coverage, multi-scale nesting, and complex geographic context. Therefore, it has been a challenging task to extract the building footprint from high resolution remote sensing images. In this study, an end-to-end Multi-Scale Geoscience Network (MS-GeoNet) is proposed for building footprint extraction. The proposed architecture focuses on multi-scale nested characteristics and the spatial correlation between buildings and surroundings. The performance of a number of embedding modules and loss functions in extracting various types of buildings are explored in detail. Our proposed method outperforms the baseline model Fully Convolutional DenseNets (FC-DenseNet) by 7.10% for the intersection over union (IoU) and by 3.09% for F1-score. Moreover, to increase the accuracy of large area interpretation, an overlap splicing and voting mechanism is proposed. It is also an effective means to solve the edge processing task. The proposed method demonstrates approximately 1.19% IoU improvement and 0.83% F1 score improvement on our dataset, compared with the traditional splicing method. MS-GeoNet is a promising approach for automatic generation of building footprint in practical applications. |
| WOS关键词 | CHINA |
| 资助项目 | Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory[GML2019ZD0301] ; Third Xinjiang Scientific Expedition[2021xjkk1303] ; National Data Sharing Infrastructure of Earth System Science |
| WOS研究方向 | Remote Sensing |
| 语种 | 英语 |
| WOS记录号 | WOS:000802165700003 |
| 出版者 | ELSEVIER |
| 资助机构 | Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory ; Third Xinjiang Scientific Expedition ; National Data Sharing Infrastructure of Earth System Science |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/178415]  |
| 专题 | 资源与环境信息系统国家重点实验室_外文论文 |
| 通讯作者 | Yao, Ling |
| 作者单位 | 1.MIT, Dept Urban Studies & Planning, Senseable City Lab, Cambridge, MA 02139 USA 2.China Univ Geosci Beijing, Beijing 100083, Peoples R China 3.Inst Geog Sci & Nat Resources Res, Chinese Acad Sci, State Key Lab Resources & Environm Informat Syst, Beijing 100101, Peoples R China 4.Southern Marine Sci & Engn Guangdong Lab, Guangzhou 511458, Peoples R China 5.Jiangsu Ctr Collaborat Innovat Geog Informat Reso, Nanjing 210023, Peoples R China 6.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, 11A, Datun Rd, Chaoyang Dist, Beijing 100101, Peoples R China |
| 推荐引用方式 GB/T 7714 | Liu, Tang,Yao, Ling,Qin, Jun,et al. Multi-scale attention integrated hierarchical networks for high-resolution building footprint extraction[J]. INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION,2022,109:11. |
| APA | Liu, Tang.,Yao, Ling.,Qin, Jun.,Lu, Ning.,Jiang, Hou.,...&Zhou, Chenghu.(2022).Multi-scale attention integrated hierarchical networks for high-resolution building footprint extraction.INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION,109,11. |
| MLA | Liu, Tang,et al."Multi-scale attention integrated hierarchical networks for high-resolution building footprint extraction".INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 109(2022):11. |
入库方式： OAI收割
来源：地理科学与资源研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。