Hütt, Christoph ORCID: 0000-0001-8381-9676 (2019). Combining Multitemporal Microwave and Optical Remote Sensing Data. Mapping of Land Use / Land Cover, Crop Type, and Crop Traits. PhD thesis, Universität zu Köln.
|
PDF
Huett2019MicrowaveOpticalRS_LULC_CropType_CropTraits.pdf - Published Version Bereitstellung unter der CC-Lizenz: Creative Commons Attribution Non-commercial No Derivatives. Download (83MB) | Preview |
Abstract
Humanity has changed the earth’s surface to a dramatic extent. This is especially true for the area used for agricultural production. Against the background of a growing world population and the associated increased demand for food, it is precisely this area that will become even more important in the future. In order not to have to allocate even more land to agricultural use, optimization and intensification is the only way out of the dilemma. In this context, precise Geoinformation of the agriculturally used area is of central importance. It is utilized for improving land use, producing yield forecasts for more stable food security, and optimizing agricultural management. Rapid developments in the field of satellite-based remote sensing sensors make it possible to monitor agricultural areas with increased spatial, spectral and temporal resolution. However, to retrieve the needed information from this data, new methods are needed. Furthermore, the quality of the data has to be verified. Only then can the presented geodata help to grow crops more sustainably and more efficiently. This thesis develops new approaches for monitoring agricultural areas using the technology of microwave remote sensing in combination with optical remote sensing and existing geodata. It is framed by the overall objective to obtain knowledge on how this combination of data can provide the necessary geoinformation for land use studies, precision farming, and agricultural monitoring systems. Hundreds of remote sensing images from more than eight different satellites were analyzed in six research studies from two different Areas of Interest (AOIs). The studies guide through various spatial scales. First, the general Land Use / Land Cover (LULC) on a regional level in a multi-sensor scenario is derived, evaluating different sensor combinations of varying resolutions. Next, an innovative method is proposed, through which the high geometric accuracy of radar-imaging satellite sensors is exploited to update the spatial accuracy of any external geodata of lower spatial accuracy. Such external data is then used in the next two studies, which focus on cost-effective crop type mapping using Synthetic Aperture Radar (SAR) images. The resulting enhanced LULC maps present the annually changing crop types of the region alongside external, official geoinformation that is not retrievable from remote sensing sensors. The last two research studies deal with a single maize field, on which high resolution optical WorldView-2 images and experimental bistatic SAR observations from TanDEM-X are assessed and combined with ground measurements. As a result, this thesis shows that, depending on the AOI and the application, different resolution demands need to be fulfilled before LULC, crop type, and crop traits mapping can be performed with adequate accuracy. The spatial resolution needs to be adapted to the particularities of the AOI. Evaluation of the sensors showed that SAR sensors proved beneficial for the study objective. Processing the SAR images is complicated, and the images are unintuitive at first sight. However, the advantage of SAR sensors is that they work even in cloudy conditions. This results in an increased temporal resolution, which is particularly important for monitoring the highly dynamic agricultural area. Furthermore, the high geometric accuracy of the SAR images proved ideal for implementing the Multi-Data Approach (MDA). Thus information-rich external geodata could be used to lower the remote sensing resolution needs, improve the accuracy of the LULC-maps, and to provide enhanced LULC-maps. The first study of the maize field demonstrates the potential of the WorldView-2 data in predicting in-field biomass variations, and its increased accuracy when fused with plant height measurements. The second study shows the potential of the TanDEM-X Constellation (TDM) to retrieve plant height from space. LULC, crop type and information on the spatial distribution of biomass can thus be derived efficiently and with high accuracy from the combination of SAR, optical satellites and external geodata. The shown analyses for acquiring such geoinformation represent a high potential for helping to solve the future challenges of agricultural production.
Item Type: | Thesis (PhD thesis) | ||||||||||||||||
Translated abstract: |
|
||||||||||||||||
Creators: |
|
||||||||||||||||
URN: | urn:nbn:de:hbz:38-97204 | ||||||||||||||||
Date: | 2019 | ||||||||||||||||
Language: | English | ||||||||||||||||
Faculty: | Faculty of Mathematics and Natural Sciences | ||||||||||||||||
Divisions: | Faculty of Mathematics and Natural Sciences > Department of Geosciences > Geographisches Institut | ||||||||||||||||
Subjects: | Data processing Computer science Natural sciences and mathematics Earth sciences Agriculture Geography and travel |
||||||||||||||||
Uncontrolled Keywords: |
|
||||||||||||||||
Date of oral exam: | 8 May 2019 | ||||||||||||||||
Referee: |
|
||||||||||||||||
Refereed: | Yes | ||||||||||||||||
URI: | http://kups.ub.uni-koeln.de/id/eprint/9720 |
Downloads
Downloads per month over past year
Export
Actions (login required)
View Item |