Domain Adaption for Fine-Grained Urban Village Extraction from Satellite Images

Qian Shi; Mengxi Liu; Xiaoping Liu; Penghua Liu; Pengyuan Zhang; Jinxing Yang; Xia Li

doi:10.1109/LGRS.2019.2947473

Domain Adaption for Fine-Grained Urban Village Extraction from Satellite Images

Qian Shi, Mengxi Liu, Xiaoping Liu, Penghua Liu, Pengyuan Zhang, Jinxing Yang, Xia Li

School of Geographic Sciences

Research output: Contribution to journal › Article › peer-review

78 Scopus citations

Abstract

Urban villages (UVs) are distinctive products formed in the process of rapid urbanization. The fine-grained mapping of UVs from satellite images has always been a considerable challenge because of the complex urban structures and the insufficiency of labeled samples. In this letter, we propose using the domain adaptation strategy to tackle the domain shift problem by employing adversarial learning to tune the semantic segmentation network so as to adaptively obtain similar outputs for input images from different domains. The proposed method was coupled with several segmentation networks, including U-Net, RefineNet, and DeepLab v3+, and the results show that domain adaptation can significantly improve the pixel-level mapping of UVs.

Original language	English
Article number	8886520
Pages (from-to)	1430-1434
Number of pages	5
Journal	IEEE Geoscience and Remote Sensing Letters
Volume	17
Issue number	8
DOIs	https://doi.org/10.1109/LGRS.2019.2947473
State	Published - Aug 2020

Keywords

Adversarial learning
domain adaptation
satellite images
semantic segmentation
urban village (UV)

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/LGRS.2019.2947473

Cite this

@article{9d490c46fd6c43eba890bbcc498209ca,

title = "Domain Adaption for Fine-Grained Urban Village Extraction from Satellite Images",

abstract = "Urban villages (UVs) are distinctive products formed in the process of rapid urbanization. The fine-grained mapping of UVs from satellite images has always been a considerable challenge because of the complex urban structures and the insufficiency of labeled samples. In this letter, we propose using the domain adaptation strategy to tackle the domain shift problem by employing adversarial learning to tune the semantic segmentation network so as to adaptively obtain similar outputs for input images from different domains. The proposed method was coupled with several segmentation networks, including U-Net, RefineNet, and DeepLab v3+, and the results show that domain adaptation can significantly improve the pixel-level mapping of UVs.",

keywords = "Adversarial learning, domain adaptation, satellite images, semantic segmentation, urban village (UV)",

author = "Qian Shi and Mengxi Liu and Xiaoping Liu and Penghua Liu and Pengyuan Zhang and Jinxing Yang and Xia Li",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.",

year = "2020",

month = aug,

doi = "10.1109/LGRS.2019.2947473",

language = "英语",

volume = "17",

pages = "1430--1434",

journal = "IEEE Geoscience and Remote Sensing Letters",

issn = "1545-598X",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "8",

}

TY - JOUR

T1 - Domain Adaption for Fine-Grained Urban Village Extraction from Satellite Images

AU - Shi, Qian

AU - Liu, Mengxi

AU - Liu, Xiaoping

AU - Liu, Penghua

AU - Zhang, Pengyuan

AU - Yang, Jinxing

AU - Li, Xia

PY - 2020/8

Y1 - 2020/8

N2 - Urban villages (UVs) are distinctive products formed in the process of rapid urbanization. The fine-grained mapping of UVs from satellite images has always been a considerable challenge because of the complex urban structures and the insufficiency of labeled samples. In this letter, we propose using the domain adaptation strategy to tackle the domain shift problem by employing adversarial learning to tune the semantic segmentation network so as to adaptively obtain similar outputs for input images from different domains. The proposed method was coupled with several segmentation networks, including U-Net, RefineNet, and DeepLab v3+, and the results show that domain adaptation can significantly improve the pixel-level mapping of UVs.

AB - Urban villages (UVs) are distinctive products formed in the process of rapid urbanization. The fine-grained mapping of UVs from satellite images has always been a considerable challenge because of the complex urban structures and the insufficiency of labeled samples. In this letter, we propose using the domain adaptation strategy to tackle the domain shift problem by employing adversarial learning to tune the semantic segmentation network so as to adaptively obtain similar outputs for input images from different domains. The proposed method was coupled with several segmentation networks, including U-Net, RefineNet, and DeepLab v3+, and the results show that domain adaptation can significantly improve the pixel-level mapping of UVs.

KW - Adversarial learning

KW - domain adaptation

KW - satellite images

KW - semantic segmentation

KW - urban village (UV)

UR - https://www.scopus.com/pages/publications/85089195139

U2 - 10.1109/LGRS.2019.2947473

DO - 10.1109/LGRS.2019.2947473

M3 - 文章

AN - SCOPUS:85089195139

SN - 1545-598X

VL - 17

SP - 1430

EP - 1434

JO - IEEE Geoscience and Remote Sensing Letters

JF - IEEE Geoscience and Remote Sensing Letters

IS - 8

M1 - 8886520

ER -

Domain Adaption for Fine-Grained Urban Village Extraction from Satellite Images

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this