Objektna klasifikacija rabe tal iz podatkov daljinskega zaznavanja

Abstract

Keywords

Data

Language:	Slovenian
Year of publishing:	2009
Source:	Ljubljana
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FGG - Faculty of Civil and Geodetic Engineering
Publisher:	[U. Kanjir]
UDC:	528.8:711.15(043.2)
COBISS:	4791905
Views:	2965
Downloads:	630
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Object oriented land use classification from remote sensing data
Secondary abstract:	Remote sensing is a cost-effective way for large area land cover/land use mapping availability. The reliability of these maps is very important for many of the Earth sciences and the development of (semi) automatic classification procedures is therefore essential. As long as pixel size remains coarser than, or at the best similar, in size to the (geographical) objects, pixel classification is the most appropriate approach. With increasing spatial resolution, and also improving spectral characteristics, of satellite images “traditional” classification methods become less effective. Therefore, object-oriented classification is being used more and more frequently. Object based classification combines contextual segmentation, which is a fundamental phase of the approach, and classification itself. Segmentation divides satellite image into a relatively homogeneous and semantically significant pixel groups (segments) based on one or more criteria of homogeneity. Beside their spectral features (colour) also shape and connection between the segments are considered in the process of the classification. Object oriented classification is used in the thesis as a tool for the identification of land use/cover in the case of two study areas in Slovenia (Kobarid, Gornja Radgona) with a different dominant cover classes.
Secondary keywords:	graduation thesis;geodesy;remote sensing;object based classification;segmentation;image analysis;land use;
File type:	application/pdf
Type (COBISS):	Undergraduate thesis
Thesis comment:	Univ. Ljubljana, Fakulteta za gradbeništvo in geodezijo
Pages:	XIV, 80 str., pril.
Type (ePrints):	thesis
Title (ePrints):	Objektna klasifikacija rabe tal iz podatkov daljinskega zaznavanja
Keywords (ePrints):	geodezija;diplomska dela;UNI;daljinsko zaznavanje;objektna klasifikacija;segmentacija;analiza podob;raba tal;
Keywords (ePrints, secondary language):	graduation thesis;geodesy;remote sensing;object based classification;segmentation;image analysis;land use;
Keywords (ePrints, secondary language):	graduation thesis;geodesy;remote sensing;object based classification;segmentation;image analysis;land use;
ID:	8308148