Uporaba satelitskih posnetkov Landsat in Sentinel za preučevanje širjenja pozidanih površin v Mestni občini Ljubljana

Škerl, Maruša (author), Krevs, Marko (mentor)

V diplomski nalogi smo zaznavali pozidane površine v Mestni občini Ljubljana s tremi različnimi sateliti med leti 2000, 2013 in 2020. Uporabili smo tri različne metode: nadzorovano klasifikacijo, spektralne indekse z mejo po Otsujevi metodi in spektralne indekse z ročno nastavljeno mejo. Namen naloge je bil torej ugotoviti, katera izmed treh preprostih metod daljinskega zaznavanja daje najboljše rezultate. Ugotovili smo, da je to nadzorovana klasifikacije, prav tako pa sta tudi Normalized Difference Vegetation Index (NDVI) in Built-up Index (BU) z ročno nastavljeno mejo prikazala dokaj dober rezultat. Najslabše rezultate je dala metoda spektralnih indeksov z mejo po Ostujevi metodi. Hkrati smo potrdili izkušnje avtorjev, ki so se ukvarjali s podobnimi tematikami, da težko primerjamo satelitske slike večih satelitov, zaradi različne kakovosti (Sabo, Corbane, Ferri, 2017).

• Language: Slovenian
• Keywords: pozidane površine, daljinsko zaznavanje, Mestna občina Ljubljana, satelitski posnetki, klasifikacija
• Work type: Bachelor thesis/paper
• Organization: FF - Faculty of Arts
• Year: 2021
• PID: 20.500.12556/RUL-130759
• Publication date in RUL: 17.09.2021

Secondary language

• Language: English
• Title: Use of Landsat and Sentinel satellite images for studying the spread of built-up areas in the Municipality of Ljubljana

Abstract

In this paper we detected built-up areas in the Municipality of Ljubljana with three different satellites between years 2000, 2013 and 2020. We used three different methods: supervised classification, spectral indices with a limit according to the Otsu method and spectral indices with a manually set limit. The purpose of the task was to determine which method gives the best results. We found that it is a supervised classification, and also Normalized Difference Vegetation Index and Built-up Index with manually set limit which showed a fairly good result. The worst results were obtained by the method of spectral indices with a limit according to the Ostuje method. At the same time, we agree with other authors which have dealt with similar topics that it is difficult to compare satellite images of multiple satellites, due to different quality (Sabo, Corbane, Ferri, 2017).

• Keywords: Built-up areas, remote sensing, Municipality of Ljubljana, satellite images, classification