A two-layer Conditional Random Field model for simultaneous classification of land cover and land use

Publikation: Beitrag in FachzeitschriftKonferenzaufsatz in FachzeitschriftForschungPeer-Review

Autorschaft

  • L. Albert
  • F. Rottensteiner
  • C. Heipke
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Seiten (von - bis)17-24
Seitenumfang8
FachzeitschriftInternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives
Jahrgang40
Ausgabenummer3
PublikationsstatusVeröffentlicht - 11 Aug. 2014
VeranstaltungISPRS Technical Commission III Symposium 2014 - Zurich, Schweiz
Dauer: 5 Sept. 20147 Sept. 2014

Abstract

This paper proposes a two-layer Conditional Random Field model for simultaneous classification of land cover and land use. Both classification tasks are integrated into a unified graphical model, which is reasonable due to the fact that land cover and land use exhibit strong contextual dependencies. In the CRF, we distinguish a land cover layer and a land use layer. Both layers differ with respect to the entities corresponding to the nodes and the classes to be distinguished. In the land cover layer, the nodes correspond to superpixels extracted from the image data, whereas in the land use layer the nodes correspond to objects of a geospatial land use database. Statistical dependencies between land cover and land use are explicitly modelled as pair-wise potentials. Thus, we obtain a consistent model, where the relations between land cover and land use are learned from representative training data. The approach is designed for input data based on aerial images. Experiments are performed on an urban test site. The experiments show the feasibility of the combination of both classification tasks into one overall approach and investigate the influence of the size of the superpixels on the classification result.

ASJC Scopus Sachgebiete

Ziele für nachhaltige Entwicklung

Zitieren

A two-layer Conditional Random Field model for simultaneous classification of land cover and land use. / Albert, L.; Rottensteiner, F.; Heipke, C.
in: International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, Jahrgang 40, Nr. 3, 11.08.2014, S. 17-24.

Publikation: Beitrag in FachzeitschriftKonferenzaufsatz in FachzeitschriftForschungPeer-Review

Albert, L, Rottensteiner, F & Heipke, C 2014, 'A two-layer Conditional Random Field model for simultaneous classification of land cover and land use', International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, Jg. 40, Nr. 3, S. 17-24. https://doi.org/10.5194/isprsarchives-XL-3-17-2014
Albert, L., Rottensteiner, F., & Heipke, C. (2014). A two-layer Conditional Random Field model for simultaneous classification of land cover and land use. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, 40(3), 17-24. https://doi.org/10.5194/isprsarchives-XL-3-17-2014
Albert L, Rottensteiner F, Heipke C. A two-layer Conditional Random Field model for simultaneous classification of land cover and land use. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives. 2014 Aug 11;40(3):17-24. doi: 10.5194/isprsarchives-XL-3-17-2014
Albert, L. ; Rottensteiner, F. ; Heipke, C. / A two-layer Conditional Random Field model for simultaneous classification of land cover and land use. in: International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives. 2014 ; Jahrgang 40, Nr. 3. S. 17-24.
Download
@article{374995df3bf74df7a2ffb7a5d8cc6e01,
title = "A two-layer Conditional Random Field model for simultaneous classification of land cover and land use",
abstract = "This paper proposes a two-layer Conditional Random Field model for simultaneous classification of land cover and land use. Both classification tasks are integrated into a unified graphical model, which is reasonable due to the fact that land cover and land use exhibit strong contextual dependencies. In the CRF, we distinguish a land cover layer and a land use layer. Both layers differ with respect to the entities corresponding to the nodes and the classes to be distinguished. In the land cover layer, the nodes correspond to superpixels extracted from the image data, whereas in the land use layer the nodes correspond to objects of a geospatial land use database. Statistical dependencies between land cover and land use are explicitly modelled as pair-wise potentials. Thus, we obtain a consistent model, where the relations between land cover and land use are learned from representative training data. The approach is designed for input data based on aerial images. Experiments are performed on an urban test site. The experiments show the feasibility of the combination of both classification tasks into one overall approach and investigate the influence of the size of the superpixels on the classification result.",
keywords = "Conditional Random Fields, Contextual classification, Land use classification, Multi-layer",
author = "L. Albert and F. Rottensteiner and C. Heipke",
year = "2014",
month = aug,
day = "11",
doi = "10.5194/isprsarchives-XL-3-17-2014",
language = "English",
volume = "40",
pages = "17--24",
number = "3",
note = "ISPRS Technical Commission III Symposium 2014 ; Conference date: 05-09-2014 Through 07-09-2014",

}

Download

TY - JOUR

T1 - A two-layer Conditional Random Field model for simultaneous classification of land cover and land use

AU - Albert, L.

AU - Rottensteiner, F.

AU - Heipke, C.

PY - 2014/8/11

Y1 - 2014/8/11

N2 - This paper proposes a two-layer Conditional Random Field model for simultaneous classification of land cover and land use. Both classification tasks are integrated into a unified graphical model, which is reasonable due to the fact that land cover and land use exhibit strong contextual dependencies. In the CRF, we distinguish a land cover layer and a land use layer. Both layers differ with respect to the entities corresponding to the nodes and the classes to be distinguished. In the land cover layer, the nodes correspond to superpixels extracted from the image data, whereas in the land use layer the nodes correspond to objects of a geospatial land use database. Statistical dependencies between land cover and land use are explicitly modelled as pair-wise potentials. Thus, we obtain a consistent model, where the relations between land cover and land use are learned from representative training data. The approach is designed for input data based on aerial images. Experiments are performed on an urban test site. The experiments show the feasibility of the combination of both classification tasks into one overall approach and investigate the influence of the size of the superpixels on the classification result.

AB - This paper proposes a two-layer Conditional Random Field model for simultaneous classification of land cover and land use. Both classification tasks are integrated into a unified graphical model, which is reasonable due to the fact that land cover and land use exhibit strong contextual dependencies. In the CRF, we distinguish a land cover layer and a land use layer. Both layers differ with respect to the entities corresponding to the nodes and the classes to be distinguished. In the land cover layer, the nodes correspond to superpixels extracted from the image data, whereas in the land use layer the nodes correspond to objects of a geospatial land use database. Statistical dependencies between land cover and land use are explicitly modelled as pair-wise potentials. Thus, we obtain a consistent model, where the relations between land cover and land use are learned from representative training data. The approach is designed for input data based on aerial images. Experiments are performed on an urban test site. The experiments show the feasibility of the combination of both classification tasks into one overall approach and investigate the influence of the size of the superpixels on the classification result.

KW - Conditional Random Fields

KW - Contextual classification

KW - Land use classification

KW - Multi-layer

UR - http://www.scopus.com/inward/record.url?scp=84924248993&partnerID=8YFLogxK

U2 - 10.5194/isprsarchives-XL-3-17-2014

DO - 10.5194/isprsarchives-XL-3-17-2014

M3 - Conference article

AN - SCOPUS:84924248993

VL - 40

SP - 17

EP - 24

JO - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives

JF - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives

SN - 1682-1750

IS - 3

T2 - ISPRS Technical Commission III Symposium 2014

Y2 - 5 September 2014 through 7 September 2014

ER -