Building Detection via Complementary Convolutional Features of Remote Sensing Images

Zeshan Lu; Kun Liu; Yongwei Zhang; Zhen Liu; Jiwen Dong; Qingjie Liu; Tao Xu

doi:10.1007/978-3-030-60633-6_53

Building Detection via Complementary Convolutional Features of Remote Sensing Images

Zeshan Lu
, Kun Liu
, Yongwei Zhang
, Zhen Liu
, Jiwen Dong
, Qingjie Liu
, Tao Xu^*

^*Corresponding author for this work

School of Computer Science and Engineering

University of Jinan
Shandong Provincial Key Laboratory of Network Based Intelligent Computing
Shandong Aerospace Electro-technology Institute

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

2 Scopus citations

Abstract

Building detection in remote sensing image plays an important role in urban planning and construction, management and military fields. Most of the building detection methods utilized deep neural networks to achieve better detection results. However, these methods either train their models on panchromatic images or on the fused images of panchromatic and multispectral images, which do not take into full consideration the advantages of high resolution and multispectral of earth observation images. In this paper, a large-scale building dataset is presented, more than 10 million single buildings were annotated both on panchromatic images and the corresponding multispectral images. Moreover, a building detection method is proposed based on complementary convolutional feature via fusion on panchromatic and multispectral images. Experimental results on the proposed dataset and ResNet-101 demonstrate that, complementary convolutional feature based building detection methods outperform the methods that only use panchromatic convolution feature and multispectral convolution feature.

Original language	English
Title of host publication	Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings
Editors	Yuxin Peng, Hongbin Zha, Qingshan Liu, Huchuan Lu, Zhenan Sun, Chenglin Liu, Xilin Chen, Jian Yang
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	638-647
Number of pages	10
ISBN (Print)	9783030606329
DOIs	https://doi.org/10.1007/978-3-030-60633-6_53
State	Published - 2020
Event	3rd Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2020 - Nanjing, China Duration: 16 Oct 2020 → 18 Oct 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12305 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	3rd Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2020
Country/Territory	China
City	Nanjing
Period	16/10/20 → 18/10/20

Keywords

Building dataset
Complementary convolutional feature
Deep neural network
Remote sensing

Access to Document

10.1007/978-3-030-60633-6_53

Cite this

Lu, Z., Liu, K., Zhang, Y., Liu, Z., Dong, J., Liu, Q., & Xu, T. (2020). Building Detection via Complementary Convolutional Features of Remote Sensing Images. In Y. Peng, H. Zha, Q. Liu, H. Lu, Z. Sun, C. Liu, X. Chen, & J. Yang (Eds.), Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings (pp. 638-647). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12305 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-60633-6_53

Lu, Zeshan ; Liu, Kun ; Zhang, Yongwei et al. / Building Detection via Complementary Convolutional Features of Remote Sensing Images. Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings. editor / Yuxin Peng ; Hongbin Zha ; Qingshan Liu ; Huchuan Lu ; Zhenan Sun ; Chenglin Liu ; Xilin Chen ; Jian Yang. Springer Science and Business Media Deutschland GmbH, 2020. pp. 638-647 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{285c3c08bed84647a7bc7f6a88d60959,

title = "Building Detection via Complementary Convolutional Features of Remote Sensing Images",

abstract = "Building detection in remote sensing image plays an important role in urban planning and construction, management and military fields. Most of the building detection methods utilized deep neural networks to achieve better detection results. However, these methods either train their models on panchromatic images or on the fused images of panchromatic and multispectral images, which do not take into full consideration the advantages of high resolution and multispectral of earth observation images. In this paper, a large-scale building dataset is presented, more than 10 million single buildings were annotated both on panchromatic images and the corresponding multispectral images. Moreover, a building detection method is proposed based on complementary convolutional feature via fusion on panchromatic and multispectral images. Experimental results on the proposed dataset and ResNet-101 demonstrate that, complementary convolutional feature based building detection methods outperform the methods that only use panchromatic convolution feature and multispectral convolution feature.",

keywords = "Building dataset, Complementary convolutional feature, Deep neural network, Remote sensing",

author = "Zeshan Lu and Kun Liu and Yongwei Zhang and Zhen Liu and Jiwen Dong and Qingjie Liu and Tao Xu",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.; 3rd Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2020 ; Conference date: 16-10-2020 Through 18-10-2020",

year = "2020",

doi = "10.1007/978-3-030-60633-6\_53",

language = "英语",

isbn = "9783030606329",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "638--647",

editor = "Yuxin Peng and Hongbin Zha and Qingshan Liu and Huchuan Lu and Zhenan Sun and Chenglin Liu and Xilin Chen and Jian Yang",

booktitle = "Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings",

address = "德国",

}

Lu, Z, Liu, K, Zhang, Y, Liu, Z, Dong, J, Liu, Q & Xu, T 2020, Building Detection via Complementary Convolutional Features of Remote Sensing Images. in Y Peng, H Zha, Q Liu, H Lu, Z Sun, C Liu, X Chen & J Yang (eds), Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12305 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 638-647, 3rd Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2020, Nanjing, China, 16/10/20. https://doi.org/10.1007/978-3-030-60633-6_53

Building Detection via Complementary Convolutional Features of Remote Sensing Images. / Lu, Zeshan; Liu, Kun; Zhang, Yongwei et al.
Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings. ed. / Yuxin Peng; Hongbin Zha; Qingshan Liu; Huchuan Lu; Zhenan Sun; Chenglin Liu; Xilin Chen; Jian Yang. Springer Science and Business Media Deutschland GmbH, 2020. p. 638-647 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12305 LNCS).

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

TY - GEN

T1 - Building Detection via Complementary Convolutional Features of Remote Sensing Images

AU - Lu, Zeshan

AU - Liu, Kun

AU - Zhang, Yongwei

AU - Liu, Zhen

AU - Dong, Jiwen

AU - Liu, Qingjie

AU - Xu, Tao

PY - 2020

Y1 - 2020

N2 - Building detection in remote sensing image plays an important role in urban planning and construction, management and military fields. Most of the building detection methods utilized deep neural networks to achieve better detection results. However, these methods either train their models on panchromatic images or on the fused images of panchromatic and multispectral images, which do not take into full consideration the advantages of high resolution and multispectral of earth observation images. In this paper, a large-scale building dataset is presented, more than 10 million single buildings were annotated both on panchromatic images and the corresponding multispectral images. Moreover, a building detection method is proposed based on complementary convolutional feature via fusion on panchromatic and multispectral images. Experimental results on the proposed dataset and ResNet-101 demonstrate that, complementary convolutional feature based building detection methods outperform the methods that only use panchromatic convolution feature and multispectral convolution feature.

AB - Building detection in remote sensing image plays an important role in urban planning and construction, management and military fields. Most of the building detection methods utilized deep neural networks to achieve better detection results. However, these methods either train their models on panchromatic images or on the fused images of panchromatic and multispectral images, which do not take into full consideration the advantages of high resolution and multispectral of earth observation images. In this paper, a large-scale building dataset is presented, more than 10 million single buildings were annotated both on panchromatic images and the corresponding multispectral images. Moreover, a building detection method is proposed based on complementary convolutional feature via fusion on panchromatic and multispectral images. Experimental results on the proposed dataset and ResNet-101 demonstrate that, complementary convolutional feature based building detection methods outperform the methods that only use panchromatic convolution feature and multispectral convolution feature.

KW - Building dataset

KW - Complementary convolutional feature

KW - Deep neural network

KW - Remote sensing

UR - https://www.scopus.com/pages/publications/85093860853

U2 - 10.1007/978-3-030-60633-6_53

DO - 10.1007/978-3-030-60633-6_53

M3 - 会议稿件

AN - SCOPUS:85093860853

SN - 9783030606329

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 638

EP - 647

BT - Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings

A2 - Peng, Yuxin

A2 - Zha, Hongbin

A2 - Liu, Qingshan

A2 - Lu, Huchuan

A2 - Sun, Zhenan

A2 - Liu, Chenglin

A2 - Chen, Xilin

A2 - Yang, Jian

PB - Springer Science and Business Media Deutschland GmbH

T2 - 3rd Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2020

Y2 - 16 October 2020 through 18 October 2020

ER -

Lu Z, Liu K, Zhang Y, Liu Z, Dong J, Liu Q et al. Building Detection via Complementary Convolutional Features of Remote Sensing Images. In Peng Y, Zha H, Liu Q, Lu H, Sun Z, Liu C, Chen X, Yang J, editors, Pattern Recognition and Computer Vision - 3rd Chinese Conference, PRCV 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 638-647. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-60633-6_53

Building Detection via Complementary Convolutional Features of Remote Sensing Images

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this