Show simple item record

Investigating the Potential of UAV-Based Low-Cost Camera Imagery for Measuring Biophysical Variables in Maize

dc.contributor.advisorSchellberg, Jürgen
dc.contributor.authorTewes, Andreas
dc.date.accessioned2020-04-24T15:42:10Z
dc.date.available2020-04-24T15:42:10Z
dc.date.issued27.11.2018
dc.identifier.urihttps://hdl.handle.net/20.500.11811/7378
dc.description.abstractThe potential for improved crop productivity is readily investigated in agronomic field experiments. Frequent measurements of biophysical crop variables are necessary to allow for confident statements on crop performance. Commonly, in-field measurements are tedious, labour-intensive, costly and spatially selective and therefore pose a challenge in field experiments. With the versatile, flexible employment of the platform and the high spatial and temporal resolution of the sensor data, Unmanned Aerial Vehicle (UAV)-based remote sensing offers the possibility to derive variables quickly, contactless and at low cost. This thesis examined if UAV-borne modified low-cost camera imagery allowed for remote estimation of the crop variables green leaf area index (gLAI) and radiation use efficiency (RUE) in a maize field trial under different management influences. For this, a field experiment was established at the university's research station Campus Klein-Altendorf southwest of Bonn in the years 2015 and 2016. In four treatments (two levels of nitrogen fertilisation and two levels of plant density) with five repetitions each, leaf growth of maize plants was supposed to occur differently. gLAI and biomass was measured destructively, UAV-based data was acquired in 14-day intervals over the entire experiment. Three studies were conducted and submitted for peer-review in international journals. In study I, three selected spectral vegetation indices (NDVI, GNDVI, 3BSI) were related to the gLAI measurements. Differing but definite relationships per treatment factor were found. gLAI estimation using the two-band indices (NDVI, GNDVI) yielded good results up to gLAI values of 3. The 3-bands approach (3BSI) did not provide improved accuracies. Comparing gLAI results to the spectral vegetation indices, it was determined that sole reliance on these was insufficient to draw the right conclusions on the impact of management factors on leaf area development in maize canopies. Study II evaluated parametric and non-parametric regression methods on their capability to estimate gLAI in maize, relying on UAV-based low-cost camera imagery with non-plants pixels (i.e. shaded and illuminated soil background) a) included in and b) excluded from the analysis. With regard to the parametric regression methods, all possible band combinations for a selected number of two- and three-band formulations as well as different fitting functions were tested. With regard to non-parametric methods, six regression algorithms (Random Forests Regression, Support Vector Regression, Relevance Vector Machines, Gaussian Process Regression, Kernel Regularized Least Squares, Extreme Learning Machine) were tested. It was found that all non-parametric methods performed better than the parametric methods, and that kernel-based algorithms outperformed the other tested algorithms. Excluding non-plant pixels from the analysis deteriorated models' performances. When using parametric regression methods, signal saturation occurred at gLAI values of about 3, and at values around 4 when employing non-parametric methods. Study III investigated if a) UAV-based low-cost camera imagery allowed estimating RUEs in different experimental plots where maize was cultivated in the growing season of 2016, b) those values were different from the ones previously reported in literature and c) there was a difference between RUEtotal and RUEgreen. Fractional cover and canopy reflectance was determined based on the RS imagery. Our study showed that RUEtotal ranges between 4.05 and 4.59, and RUEgreen between 4.11 and 4.65. These values were higher than those published in other research articles, but not outside the range of plausibility. The difference between RUEtotal and RUEgreen was minimal, possibly due to prolonged canopy greenness induced by the stay-green trait of the cultivar grown. In conclusion, UAV-based low-cost camera imagery allows for estimation of plant variables within a range of limitations.
dc.language.isoeng
dc.rightsIn Copyright
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/
dc.subjectFernerkundung
dc.subjectFlugkörper
dc.subjectKamera
dc.subjectBlattflächenindex
dc.subjectMais
dc.subject.ddc630 Landwirtschaft, Veterinärmedizin
dc.titleInvestigating the Potential of UAV-Based Low-Cost Camera Imagery for Measuring Biophysical Variables in Maize
dc.typeDissertation oder Habilitation
dc.publisher.nameUniversitäts- und Landesbibliothek Bonn
dc.publisher.locationBonn
dc.rights.accessRightsopenAccess
dc.identifier.urnhttps://nbn-resolving.org/urn:nbn:de:hbz:5n-52655
ulbbn.pubtypeErstveröffentlichung
ulbbnediss.affiliation.nameRheinische Friedrich-Wilhelms-Universität Bonn
ulbbnediss.affiliation.locationBonn
ulbbnediss.thesis.levelDissertation
ulbbnediss.dissID5265
ulbbnediss.date.accepted19.10.2018
ulbbnediss.instituteLandwirtschaftliche Fakultät : Institut für Nutzpflanzenwissenschaften und Ressourcenschutz (INRES)
ulbbnediss.fakultaetLandwirtschaftliche Fakultät
dc.contributor.coRefereeRascher, Uwe


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

The following license files are associated with this item:

InCopyright