Filtern
Erscheinungsjahr
Dokumenttyp
Schlagworte
- Fernerkundung (25) (entfernen)
Institut
- Raum- und Umweltwissenschaften (25) (entfernen)
Die Arbeit untersucht das Potential kleiner unbemannter Luftfahrtsysteme (UAS) in Landwirtschaft und Archäologie. Der Begriff UAS beinhaltet dabei: Fluggerät, Antriebsmechanismus, Sensorik, Bodenstation, Kommunikationsmittel zwischen Bodenstation und Fluggerät und weiteres Equipment. Aufgrund ihrer Flexibilität, fanden UAS seit der Jahrtausendwende eine blühende Entwicklung. Um die wachsende Weltbevölkerung zu ernähren, muss die landwirtschaftliche Produktion sensibel und nachhaltig intensiviert werden, um Nahrungssicherheit für alle zu gewährleisten und weitere Boden- und Landdegradation zu vermeiden. Präzisionslandwirtschaft umfasst technologische Verbesserungen hin zur effizienteren und weniger schädlichen landwirtschaftlichen Praxis. Hierbei ist die Verfügung über zeitnahe, leicht zugängliche hoch aufgelöste räumliche Daten eine Voraussetzung für die Nahrungsmittelproduktion. UAS schließen hier die Lücke zwischen Bodendaten und teuren bemannten Luftfahrtsysteme und selteneren Satellitenbildern. Die Vorteile der UAS-Daten liegen in der ad-hoc Akquisition großmaßstäbiger Fernerkundungsdaten, den geringeren Kosten gegenüber der bemannten Systeme und einer relativen Wetterunabhängigkeit, da auch unter Wolken geflogen werden kann. Den größten Anteil innerhalb der UAS stellen die Mini-UAS (Abfluggewicht von 5kg) und dabei vertikale Start- und Landesysteme. Diese können über Untersuchungsgebieten schweben, sind dadurch jedoch langsamer und eher geeignet für kleinere Flächen. Flugregularien und die Integration in den bemannten Luftraum werden derzeit europaweit harmonisiert und in den Mitgliedstaaten umgesetzt. Die Hauptziele dieser Arbeit lagen in der Evaluierung wie Schlüsselparametern landwirtschaftlicher Nutzpflanzen (Chlorophyll-, Stickstoffgehalt, Erntemenge, sonnendinduzierter Chlorophyll-Fluoreszenz) mittels UAS abgeleitet und wie UAS-Daten für archäologische Aufklärung genutzt werden können. Dazu wurde ein Quadrokopter (md4-1000, microdrones GmbH) mit einer digitalen Spiegelreflexkamera, einem Multispektralsensor (MiniMCA-6, Tetracam Inc.) und einer Thermalkamera (UCM, Zeiss) ausgestattet. Eine Sensitivitätsanalyse führte zur Ableitung geeigneter Wellenlängenbereiche und untersuchte bidirektionale und Flughöheneffekte auf das Multispektralsignal. Die Studie beschreibt außerdem die Vorgehensweise bei Bildaufnahme und Vorprozessierung mit besonderem Schwerpunkt auf die Multispektralkamera (530-900 nm). Die Vorprozessierung beinhaltet die Korrektur von Sensorfehlern (Linsenverzeichnung, Vignettierung, Kanalkalibrierung), die radiometrische Kalibrierung über eine empirische Korrektur mit Hilfe von Referenzspektren, Atmosphärenkorrektur und schließlich die geometrische Verarbeitung unter Verwendung von Structure from Motion Programme zur Generierung von Punktwolkenmodellen bis hin zum digitalen Orthophotomosaik und Höhenmodell in Zentimeterauflösung. In einer Weinbergsstudie (2011, 2012) wurden geeignete Beobachtungswinkel für die Untersuchung des Einflusses von Bodenbearbeitungsstrategien auf das Multispektralsignal evaluiert. Schrägichtaufnahmen von 45-° Beobachtungswinkel gegenüber Nadir waren am besten geeignet zur Ableitung pflanzenphysiolgischer Parameter und multispektraler Unterscheidung von Bodenbearbeitungstypen. So konnten Chlorophyll-Gehalte über Regressionsanalysen über mehrere saisonale Aufnahmen mit einem kreuzvalidierten R-² von 0.65, Stickstoffgehaltsindex von 0.76 (2012) und Ernte mit 0.84 (2011) und für verschiedene Zeitpunkte nach der Blüte (0.87) und während der Reifephase (0.73) ermittelt werden. Desweiteren wurde die (Fs) in einem Stickstoff-Düngung-Experiment bei Zuckerrüben von Multispektral-, Indizes und Thermaldaten untersucht (HyFlex-Kampagne 2012). Zuckerrübenvarietäten konnten spektral und thermal unterschieden werden, die Fluoreszenzindizes waren wetterbedingt, weniger erfolgreich. Außerdem konnte der Tagesgang der Fs trotz instabiler Einstrahlungsverhältnisse am Morgen abgeleitet werden. Die Werte waren jedoch gegenüber Bodenmessungen um ein Vielfaches erhöht. Archäologische Fernerkundung durch UAS wird bereits seit Jahren (z.B. mit Fesselballons) durchgeführt. Die Mustererkennung profitiert von der spektralen Ausdehnung vom menschlichen Auge hin zu multispektralen, neuerdings auch hyperspektralen Sensoren. Studien in Los Bañales, Spanien, zeigten die Möglichkeiten des Informationsgewinns durch Bildverarbeitung von UAS-Daten: vermutliche historische Siedlungsmuster konnten durch Landoberflächenklassifikation von Multispektraldaten mittels Support Vector Machines und Bestandsmusterdetektion beschrieben werden. Um qualitative hochwertige, hochaufgelöste UAS-Daten zu erhalten, sollten die Daten mit hoher Überlappung (80%) und auch Schrägsicht akquiriert und ggf. durch Referenzmessungen zur radiometrischen Kalibrierung und GPS-Messungen für geometrische Referenzierung ergänzt werden.
Considering actual climatic and land use changes the problem of available water resources or the estimation of potential flood risks gain eco-political and economical relevance. Adequate assessments, thus, require precise process-based hydrological knowledge. Spatially distributed hydrological modelling enables a both abstractive and realistic description of hydrological processes, and therefore contributes to the understanding of the hydrological system- responses. Referring to the example of the mesoscale Ruwer basin (a tributary to the Mosel river), a modified version of the distributive modelling system PRMS/MMS (Precipitation Runoff Modeling System/Modular Modeling System) is applied to calculate spatially and temporally explicit water budgets. To achieve modelling results as precise as possible, integration of detailed land use information (spatial distribution of the existing land use classes, crop- and site-specific growth patterns) is necessary. This information is derived here by analysis of multitemporal, geometrically and radiometrically pre-processed Landsat TM-data. This enables separation of different land use classes and differentiated quantification of the leaf area index (LAI). The LAI is estimated by a spectral unmixing approach using statistically optimized endmember sets, referring to the example of winter grain and grassland plots. As a result, numerical inputs (coefficients for calculating evapotranspiration, interception storages) and extracted non-numerical (classified) information can be provided for hydrological modelling. The version of PRMS applied in this study allows important land use terms to be parameterized in high temporal resolution. Using model input derived from the available satellite data, simulation results are obtained that prove to be realistic compared to gauge data and with respect to their spatial differentiation. Results differ significantly from those obtained by using parameters from literature or by experience without distinguishing specific and site-dependent growth patterns. It can be concluded that the quality of modelling results notably improves by integration and quantitative analysis of remote sensing data; thus, these methods are a significant contribution to physically-based hydrological modelling.
In past years, desertification and land degradation have been acknowledged as a major threat to human welfare world-wide, and their environmental and societal implications have sparked the formulation of the UN Convention to Combat Desertification (UNCCD). Any measure taken against desertification, or the design of dedicated early warning systems, must take into account both the spatial and temporal dimensions of process driving factors. Equally important, past and present reactions of ecosystems to physical and socio-economical disturbances or management interventions need to be understood. In this context, remote sensing and geoinformation processing support the required assessment, monitoring and modelling approaches, and hence provide an essential contribution to the scientific component of the struggle against desertification. Supported by DG Research of the European Commission, the Remote Sensing Department of the University of Trier convened RGLDD to promote scientific exchange between specialists working on the interface of remote sensing, geoinformation processing, desertification/land degradation research and its socio-economic implications. Although targeted at the scientific community, contributions with application perspectives were of crucial importance and both an overview of the current state of the art as well as operational opportunities were presented. Hosted at the Robert-Schuman Haus in Trier, the conference gained widespread attention and attracted an international audience from all parts of the world, which underlines the global dimension of land degradation and desertification processes. Based on a rigorous review of submitted abstracts, more than 100 contributions were accepted for oral and poster presentation, which are found in these proceedings edition in full paper form. Please note: This document is optimised for screen resolution, to receive a high-resolution version please contact the editors.
In order to discuss potential sustainability issues of expanding silage maize cultivation in Rhineland-Palatinate, spatially explicit monitoring is necessary. Publicly available statistical records are often not a sufficient basis for extensive research, especially on soil health, where risk factors like erosion and compaction depend on variables that are specific to every site, and hard to generalize for larger administrative aggregates. The focus of this study is to apply established classification algorithms to estimate maize abundance for each independent pixel, while at the same time accounting for their spatial relationship. Therefore, two ways to incorporate spatial autocorrelation of neighboring pixels are combined with three different classification models. The performance of each of these modeling approaches is analyzed and discussed. Finally, one prediction approach is applied to the imagery, and the overall predicted acreage is compared to publicly available data. We were able to show that Support Vector Machine (SVM) classification and Random Forests (RF) were able to distinguish maize pixels reliably, with kappa values well above 0.9 in most cases. The Generalized Linear Model (GLM) performed substantially worse. Furthermore, Regression Kriging (RK) as an approach to integrate spatial autocorrelation into the prediction model is not suitable in use cases with millions of sparsely clustered training pixels. Gaussian Blur is able to improve predictions slightly in these cases, but it is possible that this is only because it smoothes out impurities of the reference data. The overall prediction with RF classification combined with Gaussian Blur performed well, with out of bag error rates of 0.5% in 2009 and 1.3% in 2016. Despite the low error rates, there is a discrepancy between the predicted acreage and the official records, which is 20% in 2009 and 27% in 2016.
Evapotranspiration (ET) is one of the most important variables in hydrological studies. In the ET process, energy exchange and water transfer are involved. ET consists of transpiration and evaporation. The amount of plants transpiration dominates in ET. Especially in the forest regions, the ratio of transpiration to ET is in general 80-90 %. Meteorological variables, vegetation properties, precipitation and soil moisture are critical influence factors for ET generation. The study area is located in the forest area of Nahe catchment (Rhineland-Palatinate, Germany). The Nahe catchment is highly wooded. About 54.6 % of this area is covered by forest, with deciduous forest and coniferous forest are two primary types. A hydrological model, WaSiM-ETH, was employed for a long-term simulation from 1971-2003 in the Nahe catchment. In WaSiM-ETH, the potential evapotranspiration (ETP) was firstly calculated by the Penman-Monteith equation, and subsequently reduced according to the soil water content to obtain the actual evapotranspiration (ETA). The Penman-Monteith equation has been widely used and recommended for ETP estimation. The difficulties in applying this equation are the high demand of ground-measured meteorological data and the determination of surface resistance. A method combined remote sensing images with ground-measured meteorological data was also used to retrieve the ETA. This method is based on the surface properties such as surface albedo, fractional vegetation cover (FVC) and land surface temperature (LST) to obtain the latent heat flux (LE, corresponding to ETA) through the surface energy balance equation. LST is a critical variable for surface energy components estimation. It was retrieved from the TM/ETM+ thermal infrared (TIR) band. Due to the high-quality and cloudy-free requirements for TM/ETM+ data selection as well as the overlapping cycle of TM/ETM+ sensor is 16 days, images on only five dates are available during 1971-2003 (model ran) " May 15, 2000, July 05, 2001, July 19, August 04 and September 21 in 2003. It is found that the climate conditions of 2000, 2001 and 2003 are wet, medium wet and dry, respectively. Therefore, the remote sensing-retrieved observations are noncontinuous in a limited number over time but contain multiple climate conditions. Aerodynamic resistance and surface resistance are two most important parameters in the Penman-Monteith equation. However, for forest area, the aerodynamic resistance is calculated by a function of wind speed in the model. Since transpiration and evaporation are separately calculated by the Penman-Monteith equation in the model, the surface resistance was divided into canopy surface resistance rsc and soil surface resistance rse. rsc is related to the plants transpiration and rse is related to the bare soil evaporation. The interception evaporation was not taken into account due to its negligible contribution to ET rate under a dry-canopy (no rainfall) condition. Based on the remote sensing-retrieved observations, rsc and rse were calibrated in the WaSiM-ETH model for both forest types: for deciduous forest, rsc = 150 sm−1, rse = 250 sm−1; for coniferous forest, rsc = 300 sm−1, rse = 650 sm−1. We also carried out sensitivity analysis on rsc and rse. The appropriate value ranges of rsc and rse were determined as (annual maximum): for deciduous forest, [100,225] sm−1 for rsc and [50,450] sm−1 for rse; for coniferous forest, [225,375] sm−1 for rsc and [350,1200] sm−1 for rse. Due to the features of the observations that are in a limited number but contain multiple climate conditions, the statistical indices for model performance evaluation are required to be sensitive to extreme values. In this study, boxplots were found to well exhibit the model performance at both spatial and temporal scale. Nush-Sutcliffe efficiency (NSE), RMSE-observations standard deviation ratio (RSR), percent bias (PBIAS), mean bias error (MBE), mean variance of error distribution (S2d), index of agreement (d), root mean square error (RMSE) were found as appropriate statistical indices to provide additional evaluation information to the boxplots. The model performance can be judged as satisfactory if NSE > 0.5, RSR ≤ 0.7, PBIAS < -±12, MBE < -±0.45, S2d < 1.11, d > 0.79, RMSE < 0.97. rsc played a more important role than rse in ETP and ETA estimation by the Penman-Monteith equation, which is attributed to the fact that transpiration dominates in ET. The ETP estimation was found the most correlated to the relative humidity (RH), followed by air temperature (T), relative sunshine duration (SSD) and wind speed (WS). Under wet or medium wet climate conditions, ETA estimation was found the most correlated to T, followed by RH, SSD and WS. Under a water-stress condition, there were very small correlations between ETA and each meteorological variable.