HA U-Net: Improved Model for Building Extraction From High Resolution Remote Sensing Imagery
文献类型：期刊论文
| 作者 | Xu, Leilei3; Liu, Yujun4,5; Yang, Peng6,7; Chen, Hao1; Zhang, Hanyue8; Wang, Dan5; Zhang, Xin2 |
| 刊名 | IEEE ACCESS
 |
| 出版日期 | 2021 |
| 卷号 | 9页码:101972-101984 |
| 关键词 | Buildings Feature extraction Image segmentation Remote sensing Predictive models Training Task analysis Deep learning building extraction holistically-nested neural network attention mechanism weight mapping watershed algorithm |
| ISSN号 | 2169-3536 |
| DOI | 10.1109/ACCESS.2021.3097630 |
| 通讯作者 | Liu, Yujun(liuyj.20b@igsnrr.ac.cn) ; Chen, Hao(1145871257@qq.com) |
| 英文摘要 | Automatic extraction of buildings from high-resolution remote sensing images becomes an important research. Since the convolutional neural network can perform pixel-level segmentation, this technology has been applied in this field. But the increase in resolution prone to blurry segmentation because the model needs more edge detail and multi-scale detail learning. To solve this problem, a method is proposed in this paper, which consists of three parts: (1) an improved model named Holistically-Nested Attention U-Net (HA U-Net) is designed, which integrates the attention mechanism and multi-scale nested modules to supervise prediction; (2) During model training, an improved weighted loss function is proposed to make the designed model more focused on learning boundary features; (3) watershed algorithm is exploited for image post-processing to optimize segmentation results. The designed HA U-Net performs well on WHU Building Dataset and Urban3d Challenge dataset, and achieves 9.31%, 2.17% better F1-score and 10.78%, 1.77% better IOU than the standard U-Net respectively. The experimental results indicate that the proposed method can well solve the building adhesion problem. The research can serve as updating geographic databases. |
| WOS关键词 | SEGMENTATION ; FRAMEWORK ; NETWORK |
| WOS研究方向 | Computer Science ; Engineering ; Telecommunications |
| 语种 | 英语 |
| WOS记录号 | WOS:000678303900001 |
| 出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/164821]  |
| 专题 | 中国科学院地理科学与资源研究所 |
| 通讯作者 | Liu, Yujun; Chen, Hao |
| 作者单位 | 1.Tech Univ Berlin, Inst Geodesy & Geoinformat Sci, D-10553 Berlin, Germany 2.Tongji Univ, Coll Surveying & Geoinformat, Shanghai 200092, Peoples R China 3.Hohai Univ, Sch Earth Sci & Engn, Nanjing 211100, Peoples R China 4.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, Beijing 100101, Peoples R China 5.Prov Geomat Ctr Jiangsu, Nanjing 210013, Peoples R China 6.Chinese Acad Sci, Aerosp Informat Res Inst, Qilu Res Inst, Jinan 250100, Peoples R China 7.Suzhou Zhe Xin Informat Technol Co Ltd, Suzhou 215000, Peoples R China 8.Beijing Forestry Univ, Precis Forestry Key Lab Beijing, Beijing 100083, Peoples R China |
| 推荐引用方式 GB/T 7714 | Xu, Leilei,Liu, Yujun,Yang, Peng,et al. HA U-Net: Improved Model for Building Extraction From High Resolution Remote Sensing Imagery[J]. IEEE ACCESS,2021,9:101972-101984. |
| APA | Xu, Leilei.,Liu, Yujun.,Yang, Peng.,Chen, Hao.,Zhang, Hanyue.,...&Zhang, Xin.(2021).HA U-Net: Improved Model for Building Extraction From High Resolution Remote Sensing Imagery.IEEE ACCESS,9,101972-101984. |
| MLA | Xu, Leilei,et al."HA U-Net: Improved Model for Building Extraction From High Resolution Remote Sensing Imagery".IEEE ACCESS 9(2021):101972-101984. |
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。