Object-oriented land cover classification using multispectral satellite images - case study: the City of Zagreb

Valožić, Luka (2015) Object-oriented land cover classification using multispectral satellite images - case study: the City of Zagreb. Doctoral thesis, Faculty of Science > Department of Geography.

[img] PDF
Restricted to Repository staff only
Language: Croatian

Download (6MB) | Request a copy
[img] PDF (Dokumentacijska kartica, zahvale, sadržaj) - Other
Restricted to Repository staff only
Language: Croatian

Download (357kB) | Request a copy
[img] PDF (Naslovne stranice) - Cover Image
Restricted to Repository staff only
Language: Croatian

Download (225kB) | Request a copy
[img] Image (Prilog: karta, slika 3) - Supplemental Material
Restricted to Repository staff only
Language: Croatian

Download (26MB) | Request a copy
[img] Image (Prilog: karta, slika 17) - Supplemental Material
Restricted to Repository staff only
Language: Croatian

Download (26MB) | Request a copy
[img] Image (Prilog: karta, slika 18) - Supplemental Material
Restricted to Repository staff only
Language: Croatian

Download (26MB) | Request a copy
[img] Image (Prilog: karta, slika 19) - Supplemental Material
Restricted to Repository staff only
Language: Croatian

Download (26MB) | Request a copy

Abstract

An object-oriented land cover classification using RapidEye multispectral satellite imagery has been performed for the entire area of the City of Zagreb. Computer software eCognition 9.0 and ArcGIS 10 have been used for the research. Research includes both supervised and rule-based classification methods. Accuracy of classification method was assessed by error matrix and kappa statistic. Accuracy assessment resulted with an overall accuracy of classification at 92.06% and kappa statistic of 0.89. Data acquired by land cover classification were used as morphological-physiognomic characteristics for the differentiation of urban, peri-urban and rural settlements as well as for the production of dasymetric thematic map of population density. Scientific contributions made by this research are: production of land cover classification ruleset, production of land cover map and land cover shapefiles for the area of the City of Zagreb with classes: vegetation, impervious surfaces, bare soil, and water. A proposal has been made for the usage of the share of impervious surfaces as additional variable in the models for the typology of settlements. Scientific experience gained by this research will help improve higher education geography studies at the Faculty of Science, University of Zagreb.

Item Type: Thesis (Doctoral thesis)
Keywords: object-oriented classification, land cover, RapidEye multispectral imagery, urban and peri-urban area, V-I-S model
Supervisor: Toskić, Aleksandar
Date: 2015
Number of Pages: 131
Subjects: NATURAL SCIENCES > Geography
Divisions: Faculty of Science > Department of Geography
Depositing User: Inja Cahun
Date Deposited: 07 Apr 2015 11:33
Last Modified: 07 Jun 2016 11:18
URI: http://digre.pmf.unizg.hr/id/eprint/3764

Actions (login required)

View Item View Item

Nema podataka za dohvacanje citata