Unit:
Κατεύθυνση Στατιστική και Επιχειρησιακή ΈρευναLibrary of the School of Science
Author:
Oikonomidis Ioannis
Supervisors info:
Σάμης Τρέβεζας, Λέκτορας, Τμήμα Μαθηματικών, ΕΚΠΑ
Original Title:
Regression techniques for aggregate predictions of crop progress stages with Remote Sensing: An application to USA croplands
Translated title:
Regression techniques for aggregate predictions of crop progress stages with Remote Sensing: An application to USA croplands
Summary:
Precision Agriculture (PA) is a farm management plan aiming at optimizing returns on inputs whilst potentially reducing environmental impacts by using information technology, such us meteorological stations and satellites, namely Remote Sensing (RS). In order to achieve precise agricultural monitoring and prediction, a combination of advanced technology and powerful statistical methodology is considered necessary.
The first part of this master thesis aims in the development of a state-of-the-art statistical methodology for assessing and predicting spatiotemporally the phenological state of major crop species, from remote sensing data, at different scales of description (small and large agricultural areas).
In the second part, some elements from the theory of multivariate linear and multivariate logistic regression are presented, as well as an application to the problem of crop stage percentage estimation of USA croplands. In particular, the application concerns data from 2002 to 2019 for the corn fields of the Nebraska state.
%Multivariate linear models using ordinary least %squares (OLS) as well as generalized least squares %(GLS) are constructed,
The weekly Crop Progress Reports (CPRs) of the different corn phenological stages serve as the dependent variables and two remote sensing features, the Normalized Difference Vegetation Index (NDVI) and the Accumulated Growing Degree Days (AGDDs) as the independent variables. All the proposed models are
tested and compared in terms of their predictive capacity.
Main subject category:
Science
Keywords:
precision agriculture, remote sensing, NDVI, thermal time, AGDD, statistics, regression
File:
File access is restricted only to the intranet of UoA.
Thesis.pdf
3 MB
File access is restricted only to the intranet of UoA.
