Robust rooftop extraction from visible band images using higher order CRF

Er Li, John Femiani, Shibiao Xu, Xiaopeng Zhang, Peter Wonka

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

In this paper, we propose a robust framework for building extraction in visible band images. We first get an initial classification of the pixels based on an unsupervised presegmentation. Then, we develop a novel conditional random field (CRF) formulation to achieve accurate rooftops extraction, which incorporates pixel-level information and segment-level information for the identification of rooftops. Comparing with the commonly used CRF model, a higher order potential defined on segment is added in our model, by exploiting region consistency and shape feature at segment level. Our experiments show that the proposed higher order CRF model outperforms the state-of-the-art methods both at pixel and object levels on rooftops with complex structures and sizes in challenging environments.

Original languageEnglish (US)
Article number7047875
Pages (from-to)4483-4495
Number of pages13
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume53
Issue number8
DOIs
StatePublished - Aug 1 2015

Fingerprint

pixel
Pixels
experiment
Experiments
method
state of the art

Keywords

  • Buildings
  • rooftops conditional random field (CRF)
  • shadows

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Earth and Planetary Sciences(all)

Cite this

Robust rooftop extraction from visible band images using higher order CRF. / Li, Er; Femiani, John; Xu, Shibiao; Zhang, Xiaopeng; Wonka, Peter.

In: IEEE Transactions on Geoscience and Remote Sensing, Vol. 53, No. 8, 7047875, 01.08.2015, p. 4483-4495.

Research output: Contribution to journalArticle

Li, Er ; Femiani, John ; Xu, Shibiao ; Zhang, Xiaopeng ; Wonka, Peter. / Robust rooftop extraction from visible band images using higher order CRF. In: IEEE Transactions on Geoscience and Remote Sensing. 2015 ; Vol. 53, No. 8. pp. 4483-4495.
@article{68f4aa4421fe4f5ca61d0022f07290a8,
title = "Robust rooftop extraction from visible band images using higher order CRF",
abstract = "In this paper, we propose a robust framework for building extraction in visible band images. We first get an initial classification of the pixels based on an unsupervised presegmentation. Then, we develop a novel conditional random field (CRF) formulation to achieve accurate rooftops extraction, which incorporates pixel-level information and segment-level information for the identification of rooftops. Comparing with the commonly used CRF model, a higher order potential defined on segment is added in our model, by exploiting region consistency and shape feature at segment level. Our experiments show that the proposed higher order CRF model outperforms the state-of-the-art methods both at pixel and object levels on rooftops with complex structures and sizes in challenging environments.",
keywords = "Buildings, rooftops conditional random field (CRF), shadows",
author = "Er Li and John Femiani and Shibiao Xu and Xiaopeng Zhang and Peter Wonka",
year = "2015",
month = "8",
day = "1",
doi = "10.1109/TGRS.2015.2400462",
language = "English (US)",
volume = "53",
pages = "4483--4495",
journal = "IEEE Transactions on Geoscience and Remote Sensing",
issn = "0196-2892",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "8",

}

TY - JOUR

T1 - Robust rooftop extraction from visible band images using higher order CRF

AU - Li, Er

AU - Femiani, John

AU - Xu, Shibiao

AU - Zhang, Xiaopeng

AU - Wonka, Peter

PY - 2015/8/1

Y1 - 2015/8/1

N2 - In this paper, we propose a robust framework for building extraction in visible band images. We first get an initial classification of the pixels based on an unsupervised presegmentation. Then, we develop a novel conditional random field (CRF) formulation to achieve accurate rooftops extraction, which incorporates pixel-level information and segment-level information for the identification of rooftops. Comparing with the commonly used CRF model, a higher order potential defined on segment is added in our model, by exploiting region consistency and shape feature at segment level. Our experiments show that the proposed higher order CRF model outperforms the state-of-the-art methods both at pixel and object levels on rooftops with complex structures and sizes in challenging environments.

AB - In this paper, we propose a robust framework for building extraction in visible band images. We first get an initial classification of the pixels based on an unsupervised presegmentation. Then, we develop a novel conditional random field (CRF) formulation to achieve accurate rooftops extraction, which incorporates pixel-level information and segment-level information for the identification of rooftops. Comparing with the commonly used CRF model, a higher order potential defined on segment is added in our model, by exploiting region consistency and shape feature at segment level. Our experiments show that the proposed higher order CRF model outperforms the state-of-the-art methods both at pixel and object levels on rooftops with complex structures and sizes in challenging environments.

KW - Buildings

KW - rooftops conditional random field (CRF)

KW - shadows

UR - http://www.scopus.com/inward/record.url?scp=84926222218&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84926222218&partnerID=8YFLogxK

U2 - 10.1109/TGRS.2015.2400462

DO - 10.1109/TGRS.2015.2400462

M3 - Article

VL - 53

SP - 4483

EP - 4495

JO - IEEE Transactions on Geoscience and Remote Sensing

JF - IEEE Transactions on Geoscience and Remote Sensing

SN - 0196-2892

IS - 8

M1 - 7047875

ER -