Building extraction from PolSAR image based on deep CNN with polarimetric features

Xiaofang Xu; Yilong Lu; Bin Zou

doi:10.23919/IRS48640.2020.9253912

Building extraction from PolSAR image based on deep CNN with polarimetric features

Xiaofang Xu
, Yilong Lu
, Bin Zou

School of Electronics and Information Engineering, Harbin Institute of Technology
Nanyang Technological University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

For polarimetric synthetic aperture radar (PolSAR) images, building extraction has been a challenging topic for long time in applications of land-use and land-cover analysis. Due to similar structures of buildings and such vegetation as forest, they often exhibit similar PolSAR scattering characteristics that are often difficult to distinguishing. Recently, deep Convolutional Neural Network (CNN) has been widely investigated for image processing with many promising results. This paper proposes a method that combines polarimetric features with the CNN network to realize the comprehensive utilization of polarimetric and contextual information of PolSAR data for the extraction of building areas in PolSAR images. Comparison experiments on both ESAR and EMISAR L-band PolSAR datasets show that the proposed method can generate better results for building extraction.

Original language	English
Title of host publication	2020 21st International Radar Symposium, IRS 2020
Publisher	IEEE Computer Society
Pages	117-120
Number of pages	4
ISBN (Electronic)	9788394942151
DOIs	https://doi.org/10.23919/IRS48640.2020.9253912
State	Published - 5 Oct 2020
Externally published	Yes
Event	21st International Radar Symposium, IRS 2020 - Warsaw, Poland Duration: 5 Oct 2020 → 7 Oct 2020

Publication series

Name	Proceedings International Radar Symposium
Volume	2020-October
ISSN (Print)	2155-5753

Conference

Conference	21st International Radar Symposium, IRS 2020
Country/Territory	Poland
City	Warsaw
Period	5/10/20 → 7/10/20

UN SDGs

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

SDG 15 Life on Land

Keywords

Building extraction
Convolutional neural network (CNN)
PolSAR
Polarimetric features
SAR

Access to Document

10.23919/IRS48640.2020.9253912

Cite this

@inproceedings{d48ea5d4928b4647a6175ac4f2c03bb6,

title = "Building extraction from PolSAR image based on deep CNN with polarimetric features",

abstract = "For polarimetric synthetic aperture radar (PolSAR) images, building extraction has been a challenging topic for long time in applications of land-use and land-cover analysis. Due to similar structures of buildings and such vegetation as forest, they often exhibit similar PolSAR scattering characteristics that are often difficult to distinguishing. Recently, deep Convolutional Neural Network (CNN) has been widely investigated for image processing with many promising results. This paper proposes a method that combines polarimetric features with the CNN network to realize the comprehensive utilization of polarimetric and contextual information of PolSAR data for the extraction of building areas in PolSAR images. Comparison experiments on both ESAR and EMISAR L-band PolSAR datasets show that the proposed method can generate better results for building extraction.",

keywords = "Building extraction, Convolutional neural network (CNN), PolSAR, Polarimetric features, SAR",

author = "Xiaofang Xu and Yilong Lu and Bin Zou",

note = "Publisher Copyright: {\textcopyright} 2020 Warsaw University of Technology.; 21st International Radar Symposium, IRS 2020 ; Conference date: 05-10-2020 Through 07-10-2020",

year = "2020",

month = oct,

day = "5",

doi = "10.23919/IRS48640.2020.9253912",

language = "英语",

series = "Proceedings International Radar Symposium",

publisher = "IEEE Computer Society",

pages = "117--120",

booktitle = "2020 21st International Radar Symposium, IRS 2020",

address = "美国",

}

Xu, X, Lu, Y & Zou, B 2020, Building extraction from PolSAR image based on deep CNN with polarimetric features. in 2020 21st International Radar Symposium, IRS 2020., 9253912, Proceedings International Radar Symposium, vol. 2020-October, IEEE Computer Society, pp. 117-120, 21st International Radar Symposium, IRS 2020, Warsaw, Poland, 5/10/20. https://doi.org/10.23919/IRS48640.2020.9253912

Building extraction from PolSAR image based on deep CNN with polarimetric features. / Xu, Xiaofang; Lu, Yilong; Zou, Bin.
2020 21st International Radar Symposium, IRS 2020. IEEE Computer Society, 2020. p. 117-120 9253912 (Proceedings International Radar Symposium; Vol. 2020-October).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Building extraction from PolSAR image based on deep CNN with polarimetric features

AU - Xu, Xiaofang

AU - Lu, Yilong

AU - Zou, Bin

PY - 2020/10/5

Y1 - 2020/10/5

N2 - For polarimetric synthetic aperture radar (PolSAR) images, building extraction has been a challenging topic for long time in applications of land-use and land-cover analysis. Due to similar structures of buildings and such vegetation as forest, they often exhibit similar PolSAR scattering characteristics that are often difficult to distinguishing. Recently, deep Convolutional Neural Network (CNN) has been widely investigated for image processing with many promising results. This paper proposes a method that combines polarimetric features with the CNN network to realize the comprehensive utilization of polarimetric and contextual information of PolSAR data for the extraction of building areas in PolSAR images. Comparison experiments on both ESAR and EMISAR L-band PolSAR datasets show that the proposed method can generate better results for building extraction.

AB - For polarimetric synthetic aperture radar (PolSAR) images, building extraction has been a challenging topic for long time in applications of land-use and land-cover analysis. Due to similar structures of buildings and such vegetation as forest, they often exhibit similar PolSAR scattering characteristics that are often difficult to distinguishing. Recently, deep Convolutional Neural Network (CNN) has been widely investigated for image processing with many promising results. This paper proposes a method that combines polarimetric features with the CNN network to realize the comprehensive utilization of polarimetric and contextual information of PolSAR data for the extraction of building areas in PolSAR images. Comparison experiments on both ESAR and EMISAR L-band PolSAR datasets show that the proposed method can generate better results for building extraction.

KW - Building extraction

KW - Convolutional neural network (CNN)

KW - PolSAR

KW - Polarimetric features

KW - SAR

UR - https://www.scopus.com/pages/publications/85097264490

U2 - 10.23919/IRS48640.2020.9253912

DO - 10.23919/IRS48640.2020.9253912

M3 - 会议稿件

AN - SCOPUS:85097264490

T3 - Proceedings International Radar Symposium

SP - 117

EP - 120

BT - 2020 21st International Radar Symposium, IRS 2020

PB - IEEE Computer Society

T2 - 21st International Radar Symposium, IRS 2020

Y2 - 5 October 2020 through 7 October 2020

ER -

Building extraction from PolSAR image based on deep CNN with polarimetric features

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this