Cross-sensors analysis of multispectral Earth observation data for environmental monitoring

Trevisiol, Francesca (2023) Cross-sensors analysis of multispectral Earth observation data for environmental monitoring, [Dissertation thesis], Alma Mater Studiorum Università di Bologna. Dottorato di ricerca in Ingegneria civile, chimica, ambientale e dei materiali, 35 Ciclo. DOI 10.48676/unibo/amsdottorato/11038.
Documenti full-text disponibili:
[img] Documento PDF (English) - Richiede un lettore di PDF come Xpdf o Adobe Acrobat Reader
Disponibile con Licenza: Creative Commons Attribution Non-commercial No Derivatives 4.0 (CC BY-NC-ND 4.0) .
Download (8MB)


Time Series Analysis of multispectral satellite data offers an innovative way to extract valuable information of our changing planet. This is now a real option for scientists thanks to data availability as well as innovative cloud-computing platforms, such as Google Earth Engine. The integration of different missions would mitigate known issues in multispectral time series construction, such as gaps due to clouds or other atmospheric effects. With this purpose, harmonization among Landsat-like missions is possible through statistical analysis. This research offers an overview of the different instruments from Landsat and Sentinel missions (TM, ETM, OLI, OLI-2 and MSI sensors) and products levels (Collection-2 Level-1 and Surface Reflectance for Landsat and Level-1C and Level-2A for Sentinel-2). Moreover, a cross-sensors comparison was performed to assess the interoperability of the sensors on-board Landsat and Sentinel-2 constellations, having in mind a possible combined use for time series analysis. Firstly, more than 20,000 pairs of images almost simultaneously acquired all over Europe were selected over a period of several years. The study performed a cross-comparison analysis on these data, and provided an assessment of the calibration coefficients that can be used to minimize differences in the combined use. Four of the most popular vegetation indexes were selected for the study: NDVI, EVI, SAVI and NDMI. As a result, it is possible to reconstruct a longer and denser harmonized time series since 1984, useful for vegetation monitoring purposes. Secondly, the spectral characteristics of the recent Landsat-9 mission were assessed for a combined use with Landsat-8 and Sentinel-2. A cross-sensor analysis of common bands of more than 3,000 almost simultaneous acquisitions verified a high consistency between datasets. The most relevant discrepancy has been observed in the blue and SWIRS bands, often used in vegetation and water related studies. This analysis was supported with spectroradiometer ground measurements.

Tipologia del documento
Tesi di dottorato
Trevisiol, Francesca
Dottorato di ricerca
Settore disciplinare
Settore concorsuale
Parole chiave
satellite remote sensing, Landsat, Sentinel, multispectral, harmonization, time-series
Data di discussione
15 Giugno 2023

Altri metadati

Statistica sui download

Gestione del documento: Visualizza la tesi