ALBERT

Description: Knowledge of the location and extent of agricultural fields is required for many applications, including agricultural statistics, environmental monitoring, and administrative policies. Furthermore, many mapping applications, such as object-based classification, crop type distinction, or large-scale yield prediction benefit significantly from the accurate delineation of fields. Still, most existing field maps and observation systems rely on historic administrative maps or labor-intensive field campaigns. These are often expensive to maintain and quickly become outdated, especially in regions of frequently changing agricultural patterns. However, exploiting openly available remote sensing imagery (e.g., from the European Union’s Copernicus programme) may allow for frequent and efficient field mapping with minimal human interaction. We present a new approach to extracting agricultural fields at the sub-pixel level. It consists of boundary detection and a field polygon extraction step based on a newly developed, modified version of the growing snakes active contours model we refer to as graph-based growing contours. This technique is capable of extracting complex networks of boundaries present in agricultural landscapes, and is largely automatic with little supervision required. The whole detection and extraction process is designed to work independently of sensor type, resolution, or wavelength. As a test case, we applied the method to two regions of interest in a study area in the northern Germany using multi-temporal Sentinel-2 imagery. Extracted fields were compared visually and quantitatively to ground reference data. The technique proved reliable in producing polygons closely matching reference data, both in terms of boundary location and statistical proxies such as median field size and total acreage.

Description: In China, freshwater is an increasingly scarce resource and wetlands are under great pressure. This study focuses on China’s second largest freshwater lake in the middle reaches of the Yangtze River—the Dongting Lake—and its surrounding wetlands, which are declared a protected Ramsar site. The Dongting Lake area is also a research region of focus within the Sino-European Dragon Programme, aiming for the international collaboration of Earth Observation researchers. ESA’s Copernicus Programme enables comprehensive monitoring with area-wide coverage, which is especially advantageous for large wetlands that are difficult to access during floods. The first year completely covered by Sentinel-1 SAR satellite data was 2016, which is used here to focus on Dongting Lake’s wetland dynamics. The well-established, threshold-based approach and the high spatio-temporal resolution of Sentinel-1 imagery enabled the generation of monthly surface water maps and the analysis of the inundation frequency at a 10 m resolution. The maximum extent of the Dongting Lake derived from Sentinel-1 occurred in July 2016, at 2465 km2, indicating an extreme flood year. The minimum size of the lake was detected in October, at 1331 km2. Time series analysis reveals detailed inundation patterns and small-scale structures within the lake that were not known from previous studies. Sentinel-1 also proves to be capable of mapping the wetland management practices for Dongting Lake polders and dykes. For validation, the lake extent and inundation duration derived from the Sentinel-1 data were compared with excerpts from the Global WaterPack (frequently derived by the German Aerospace Center, DLR), high-resolution optical data, and in situ water level data, which showed very good agreement for the period studied. The mean monthly extent of the lake in 2016 from Sentinel-1 was 1798 km2, which is consistent with the Global WaterPack, deviating by only 4%. In summary, the presented analysis of the complete annual time series of the Sentinel-1 data provides information on the monthly behavior of water expansion, which is of interest and relevance to local authorities involved in water resource management tasks in the region, as well as to wetland conservationists concerned with the Ramsar site wetlands of Dongting Lake and to local researchers.

Description: Field mapping and information on agricultural landscapes is of increasing importance for many applications. Monitoring schemes and national cadasters provide a rich source of information but their maintenance and regular updating is costly and labor-intensive. Automatized mapping of fields based on remote sensing imagery may aid in this task and allow for a faster and more regular observation. Although remote sensing has seen extensive use in agricultural research topics, such as plant health monitoring, crop type classification, yield prediction, and irrigation, field delineation and extraction has seen comparatively little research interest. In this study, we present a field boundary detection technique based on deep learning and a variety of image features, and combine it with the graph-based growing contours (GGC) method to extract agricultural fields in a study area in northern Germany. The boundary detection step only requires red, green, and blue (RGB) data and is therefore largely independent of the sensor used. We compare different image features based on color and luminosity information and evaluate their usefulness for the task of field boundary detection. A model based on texture metrics, gradient information, Hessian matrix eigenvalues, and local statistics showed good results with accuracies up to 88.2%, an area under the ROC curve (AUC) of up to 0.94, and F1 score of up to 0.88. The exclusive use of these universal image features may also facilitate transferability to other regions. We further present modifications to the GGC method intended to aid in upscaling of the method through process acceleration with a minimal effect on results. We combined the boundary detection results with the GGC method for field polygon extraction. Results were promising, with the new GGC version performing similarly or better than the original version while experiencing an acceleration of 1.3× to 2.3× on different subsets and input complexities. Further research may explore other applications of the GGC method outside agricultural remote sensing and field extraction.

Description: A rapid increase in the world’s population over the last century has triggered the transformation of the earth surface, especially in urban areas, where more than half of the global population live. Ghana is no exception and a high population growth rate, coupled with economic development over the last three decades, has transformed the Greater Accra region into a hotspot for massive urban growth. The urban extent of the region has expanded extensively, mainly at the expense of the vegetative cover in the region. Although urbanization presents several opportunities, the environmental and social problems cannot be underestimated. Therefore, the need to estimate the rate and extent of land use/land cover changes in the region and the main drivers of these changes is imperative. Geographic Information Systems (GIS) and remote sensing techniques provide effective tools in studying and monitoring land-use/land-cover change over space and time. A post classification change detection of multiple Landsat images was conducted to map and analyse the extent and rate of land use/land cover change in the region between 1991 and 2015. Subsequently, the urban extent of the region was forecasted for the year 2025 using the Markov Chain and the Multi-Layer Perceptron neural network, together with drivers representing proximity, biophysical, and socio-economic variables. The results from the research revealed that built-up areas increased by 277% over the 24-year study period. However, forest areas experienced massive reduction, diminishing from 34% in 1991 to 6.5% in 2015. The 2025 projected land use map revealed that the urban extent will massively increase to cover 70% of the study area, as compared to 44% in 2015. The urban extent is also anticipated to spill into the adjoining districts mainly on the western and eastern sides of the region. The success of this research in generating a future land-use map for 2025, together with the other significant findings, demonstrates the usefulness of spatial models as tools for sustainable city planning and environmental management, especially for urban planners in developing countries.

Description: Melt ponds are key elements in the energy balance of Arctic sea ice. Observing their temporal evolution is crucial for understanding melt processes and predicting sea ice evolution. Remote sensing is the only technique that enables large-scale observations of Arctic sea ice. However, monitoring melt pond deepening in this way is challenging because most of the optical signal reflected by a pond is defined by the scattering characteristics of the underlying ice. Without knowing the influence of meltwater on the reflected signal, the water depth cannot be determined. To solve the problem, we simulated the way meltwater changes the reflected spectra of bare ice. We developed a model based on the slope of the log-scaled remote sensing reflectance at 710 nm as a function of depth that is widely independent from the bottom albedo and accounts for the influence of varying solar zenith angles. We validated the model using 49 in situ melt pond spectra and corresponding depths from shallow ponds on dark and bright ice. Retrieved pond depths are accurate (root mean square error, RMSE=2.81 cm; nRMSE=16 %) and highly correlated with in situ measurements (r=0.89; p=4.34×10-17). The model further explains a large portion of the variation in pond depth (R2=0.74). Our results indicate that our model enables the accurate retrieval of pond depth on Arctic sea ice from optical data under clear sky conditions without having to consider pond bottom albedo. This technique is potentially transferrable to hyperspectral remote sensors on unmanned aerial vehicles, aircraft and satellites.

Description: Hyperspectral remote-sensing instruments on unmanned aerial vehicles (UAVs), aircraft and satellites offer new opportunities for sea ice observations. We present the first study using airborne hyperspectral imagery of Arctic sea ice and evaluate two atmospheric correction approaches (ATCOR-4 (Atmospheric and Topographic Correction version 4; v7.0.0) and empirical line calibration). We apply an existing, field data-based model to derive the depth of melt ponds, to airborne hyperspectral AisaEAGLE imagery and validate results with in situ measurements. ATCOR-4 results roughly match the shape of field spectra but overestimate reflectance resulting in high root-mean-square error (RMSE) (between 0.08 and 0.16). Noisy reflectance spectra may be attributed to the low flight altitude of 200 ft and Arctic atmospheric conditions. Empirical line calibration resulted in smooth, accurate spectra (RMSE 〈 0.05) that enabled the assessment of melt pond bathymetry. Measured and modeled pond bathymetry are highly correlated (r = 0.86) and accurate (RMSE = 4.04 cm), and the model explains a large portion of the variability (R2 = 0.74). We conclude that an accurate assessment of melt pond bathymetry using airborne hyperspectral data is possible subject to accurate atmospheric correction. Furthermore, we see the necessity to improve existing approaches with Arctic-specific atmospheric profiles and aerosol models and/or by using multiple reference targets on the ground.

Description: Submerged aquatic vegetation (SAV) plays an important role in freshwater lake ecosystems. Due to its sensitivity to environmental changes, several SAV species serve as bioindicators for the trophic state of freshwater lakes. Variations in water temperature, light availability and nutrient concentration affect SAV growth and species composition. To monitor the trophic state as required by the European Water Framework Directive (WFD), SAV needs to be monitored regularly. This study analyses the development of macrophyte patches at Lake Starnberg, Germany, by exploring four Sentinel-2A acquired within the main growing season in August and September 2015. Two different methods of littoral bottom coverage assessment are compared, i.e. a semi-empirical method using depth-invariant indices and a physically based, bio-optical method using WASI-2D (Water Colour Simulator). For a precise Sentinel-2 imaging by date and hour, satellite measurements were supported by lake bottom spectra delivered by in situ data based reflectance models. Both methods identified vegetated and non-vegetated patches in shallow water areas. Furthermore, tall- and meadow-growing SAV growth classes could be differentiated. Both methods revealed similar results when focusing on the identification of sediment and SAV patches (R² from 0.56 to 0.81), but not for a differentiation on SAV class growth level (R²