Geodetski list, Vol. 71 (94) No. 1, 2017.
Izvorni znanstveni članak
Geostatistical Analysis of Vegetation Indices in Forest Ecosystem Česma
Ela Vela
; Geodetski fakultet Sveučilišta u Zagrebu, Zagreb, Hrvatska
Ivan Medved
; Hrvatski šumarski institut, Jastrebarsko, Hrvatska
Vanja Miljković
orcid.org/0000-0003-2290-3844
; Geodetski fakultet Sveučilišta u Zagrebu, Zagreb, Hrvatska
Sažetak
Because of a possibility to gather and analyse a large amount of spatial data in the longer temporal period, we can get new information which alleviates the implementation of geostatistical analysis on a larger area. Multispectral satellite images with spatial resolution of 30 m taken with Landsat 7 and Landsat 8 satellites were being used in the representative smoothing in the geostatistical analysis. In this study the geostatistical analysis NDVI, NDWI, GNDVI, EVI and SAVI was conducted through a ten-year period in the piezometer points in the Česma forest. This study assesses vegetation indices calculated from the multispectral images on the basis of absorption, transmission and reflection of energy from the vegetation in different spectral bands. Vegetation indices consistently provide spatio-temporal information about vegetation cover, and in this study the forest composition is being observed in two ways. Foundation for the first one are the differences based upon calculated vegetation indices from the raw data where the value of a pixel is DN, and foundation for the second one are the atmospheric corrected satellite images with DOS1 correction. Automatization of data processing was made in statistical program language R which accelerated the processing of a large amount of raster images. R was used for the implementation of the algorithm and its purpose was to conduct a quality geostatistical analysis.
Ključne riječi
geostatistics; vegetation indices; atmospheric correction DOS1; correlation; spatio-temporal analysis
Hrčak ID:
181953
URI
Datum izdavanja:
27.3.2017.
Posjeta: 2.230 *