A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF
BRNO (CZECHIA).

J 2015

A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF BRNO (CZECHIA).

KÝNOVÁ, Andrea and Petr DOBROVOLNÝ

Basic information

Original name

A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF BRNO (CZECHIA).

Authors

KÝNOVÁ, Andrea (203 Czech Republic, belonging to the institution) and Petr DOBROVOLNÝ (203 Czech Republic, belonging to the institution)

Edition

Acta Universitatis Carolinae Geographica, 2015, 0300-5402

Other information

Language

English

Type of outcome

Článek v odborném periodiku

Field of Study

Earth magnetism, geodesy, geography

Country of publisher

Czech Republic

Confidentiality degree

není předmětem státního či obchodního tajemství

References:

URL

RIV identification code

RIV/00216224:14310/15:00085925

Organization unit

Faculty of Science

DOI

http://dx.doi.org/10.14712/23361980.2015.94

Keywords in English

image classification; multilayer perceptron; urban land cover; ASTER

Abstract

V originále

Accurate and updated land cover maps provide crucial basic information in a number of important enterprises, with sustainable development and regional planning far from the least of them. Remote sensing is probably the most efficient approach to obtaining a land cover map. However, certain intrinsic limitations limit the accuracy of automatic approaches to image classification. Classifications within highly heterogeneous urban areas are especially challenging. This study makes a presentation of multilayer perceptron (MLP), an artificial neural network (ANN), as an applicable approach to image classification. Optimal MLP architecture parameters were established by means of a training set. The resulting network was used to classify a sub-scene within ASTER imagery. The results were evaluated against a test dataset. The overall accuracy of classification was 94.8%. This is comparable to classification results from a maximum likelihood classifier (MLC) used for the same image. In built-up areas, MLP did not exaggerate built-up areas at the expense of other classes to the same extent as MLC.

Links

MUNI/A/0952/2013, interní kód MU

Name: Analýza, hodnocení a vizualizace globálních environmentálních změn v krajinné sféře Země (Acronym: AVIGLEZ)

Investor: Masaryk University, Category A

Citovat

KÝNOVÁ, Andrea and Petr DOBROVOLNÝ. A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF BRNO (CZECHIA). Acta Universitatis Carolinae Geographica. 2015, vol. 50, No 2, p. 153-163. ISSN 0300-5402. Available from: https://dx.doi.org/10.14712/23361980.2015.94.

@article{1324278,
   author = {Kýnová, Andrea and Dobrovolný, Petr},
   article_number = {2},
   doi = {http://dx.doi.org/10.14712/23361980.2015.94},
   keywords = {image classification; multilayer perceptron; urban land cover; ASTER},
   language = {eng},
   issn = {0300-5402},
   journal = {Acta Universitatis Carolinae Geographica},
   title = {A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF BRNO (CZECHIA).},
   url = {http://www.aucgeographica.cz/index.php/AUC_Geographica/article/view/98/pdf_56},
   volume = {50},
   year = {2015}
}

TY  - JOUR
ID  - 1324278
AU  - Kýnová, Andrea - Dobrovolný, Petr
PY  - 2015
TI  - A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF BRNO (CZECHIA).
JF  - Acta Universitatis Carolinae Geographica
VL  - 50
IS  - 2
SP  - 153-163
EP  - 153-163
SN  - 03005402
KW  - image classification
KW  - multilayer perceptron
KW  - urban land cover
KW  - ASTER
UR  - http://www.aucgeographica.cz/index.php/AUC_Geographica/article/view/98/pdf_56
L2  - http://www.aucgeographica.cz/index.php/AUC_Geographica/article/view/98/pdf_56
N2  - Accurate and updated land cover maps provide crucial basic information in a number of important enterprises, with sustainable development and regional planning far from the least of them. Remote sensing is probably the most efficient approach to obtaining a land cover map. However, certain intrinsic limitations limit the accuracy of automatic approaches to image classification. Classifications within highly heterogeneous urban areas are especially challenging. This study makes a presentation of multilayer perceptron (MLP), an artificial neural network (ANN), as an applicable approach to image classification. Optimal MLP architecture parameters were established by means of a training set. The resulting network was used to classify a sub-scene within ASTER imagery. The results were evaluated against a test dataset. The overall accuracy of classification was 94.8%. This is comparable to classification results from a maximum likelihood classifier (MLC) used for the same image. In built-up areas, MLP did not exaggerate built-up areas at the expense of other classes to the same extent as MLC.
ER  -

KÝNOVÁ, Andrea and Petr DOBROVOLNÝ. A NEURAL NETS URBAN LAND COVER CLASSIFICATION: A CASE STUDY OF BRNO (CZECHIA). \textit{Acta Universitatis Carolinae Geographica}. 2015, vol.~50, No~2, p.~153-163. ISSN~0300-5402. Available from: https://dx.doi.org/10.14712/23361980.2015.94.

Detailed Information on Publication Record