Explore the vast range of titles available.

In this study, we further develop the processing of ground-based interferometric radar measurements for the application of bridge monitoring. Applying ground-based radar in such complex setups or long measurement durations requires advanced processing steps to receive accurate measurements. These steps involve removing external influences from the measurement and evaluating the measurement uncertainty during processing. External influences include disturbances caused by objects moving through the signal, static clutter from additional scatterers, and changes in atmospheric properties. After removing these influences, the line-of-sight displacement vectors, measured by multiple ground-based radars, are decomposed into three-dimensional displacement components. The advanced processing steps are applied exemplarily on measurements with two sensors at a prestressed concrete bridge near Coburg (Germany). The external influences are successfully removed, and two components of the three-dimensional displacement vector are determined. A measurement uncertainty of less than 0.1 mm is achieved for the discussed application.

Land cover and its change are crucial for many environmental applications. This study focuses on the land cover classification and change detection with multitemporal and multispectral Sentinel-2 satellite data. To address the challenging land cover change detection task, we rely on two different deep learning architectures and selected pre-processing steps. For example, we define an excluded class and deal with temporal water shoreline changes in the pre-processing. We employ a fully convolutional neural network (FCN), and we combine the FCN with long short-term memory (LSTM) networks. The FCN can only handle monotemporal input data, while the FCN combined with LSTM can use sequential information (multitemporal). Besides, we provided fixed and variable sequences as training sequences for the combined FCN and LSTM approach. The former refers to using six defined satellite images, while the latter consists of image sequences from an extended training pool of ten images. Further, we propose measures for the robustness concerning the selection of Sentinel-2 image data as evaluation metrics. We can distinguish between actual land cover changes and misclassifications of the deep learning approaches with these metrics. According to the provided metrics, both multitemporal LSTM approaches outperform the monotemporal FCN approach, about 3 to 5 percentage points (p.p.). The LSTM approach trained on the variable sequences detects 3 p.p. more land cover changes than the LSTM approach trained on the fixed sequences. Besides, applying our selected pre-processing improves the water classification and avoids reducing the dataset effectively by 17.6%. The presented LSTM approaches can be modified to provide applicability for a variable number of image sequences since we published the code of the deep learning models. The Sentinel-2 data and the ground truth are also freely available.

Bushfires pose a severe risk, among others, to humans, wildlife, and infrastructures. Rapid detection of fires is crucial for fire-extinguishing activities and rescue missions. Besides, mapping burned areas also supports evacuation and accessibility to emergency facilities. In this study, we propose a generic approach for detecting fires and burned areas based on machine learning (ML) approaches and remote sensing data. While most studies investigated either the detection of fires or mapping burned areas, we addressed and evaluated, in particular, the combined detection on three selected case study regions. Multispectral Sentinel-2 images represent the input data for the supervised ML models. First, we generated the reference data for the three target classes, burned, unburned, and fire, since no reference data were available. Second, the three regional fire datasets were preprocessed and divided into training, validation, and test subsets according to a defined schema. Furthermore, an undersampling approach ensured the balancing of the datasets. Third, seven selected supervised classification approaches were used and evaluated, including tree-based models, a self-organizing map, an artificial neural network, and a one-dimensional convolutional neural network (1D-CNN). All selected ML approaches achieved satisfying classification results. Moreover, they performed a highly accurate fire detection, while separating burned and unburned areas was slightly more challenging. The 1D-CNN and extremely randomized tree were the best-performing models with an overall accuracy score of 98% on the test subsets. Even on an unknown test dataset, the 1D-CNN achieved high classification accuracies. This generalization is even more valuable for any use-case scenario, including the organization of fire-fighting activities or civil protection. The proposed combined detection could be extended and enhanced with crowdsourced data in further studies.

Machine learning approaches are valuable methods in hyperspectral remote sensing, especially for the classification of land cover or for the regression of physical parameters. While the recording of hyperspectral data has become affordable with innovative technologies, the acquisition of reference data (ground truth) has remained expensive and time-consuming. There is a need for methodological approaches that can handle datasets with significantly more hyperspectral input data than reference data. We introduce the Supervised Self-organizing Maps (SuSi) framework, which can perform unsupervised, supervised and semi-supervised classification as well as regression on high-dimensional data. The methodology of the SuSi framework is presented and compared to other frameworks. Its different parts are evaluated on two hyperspectral datasets. The results of the evaluations can be summarized in four major findings: (1) The supervised and semi-Supervised Self-organizing Maps (SOM) outperform random forest in the regression of soil moisture. (2) In the classification of land cover, the supervised and semi-supervised SOM reveal great potential. (3) The unsupervised SOM is a valuable tool to understand the data. (4) The SuSi framework is versatile, flexible, and easy to use. The SuSi framework is provided as an open-source Python package on GitHub.

Information about the chlorophyll a concentration of inland water bodies is essential for water monitoring. This study focuses on estimating chlorophyll a with remote sensing data, and machine learning (ML) approaches on the real-world SpecWa dataset. We adapt and apply a one-dimensional convolutional neural network (1D CNN) as a deep learning architecture for the first time to address this estimation. Since such a DL approach requires a large amount of data for its training, we rely on simulation data generated by the Water Color Simulator (WASI). This simulation is prepared accordingly and includes a knowledge-based water composition with two origins of the chlorophyll a concentration. Therefore, the training data is independent of the real-world SpecWa dataset, which is challenging for any ML approach. We define two spectral downsampling approaches as a pre-processing step, representing the hyperspectral EnMAP satellite mission (SR-EnMAP) and the multispectral Sentinel-2 mission (SR-Sentinel). Subsequently, we train a Random Forest, an artificial neural network, a band-ratio approach, and the 1D CNN on the WASI-generated simulation training dataset. Finally, all ML models are evaluated on the real SpecWa dataset. For both downsampled data, the 1D CNN outperforms the other ML models. On the finer resolved SR-EnMAP data it achieves an R2=81.9%, RMSE=12.4 μg L−1, and MAE=6.7 μg L−1. Besides, the 1D CNN’s performance decreases on the SR-Sentinel data to R2=62.4%. When focusing on the individual water bodies of the SpecWa dataset, the most significant differences exist between natural and artificial water bodies. We discover that the applied models estimate the chlorophyll a concentration of most natural water bodies satisfyingly. In sum, the newly DL approach can estimate the chlorophyll a values of unknown inland water bodies successfully, although it is trained on an entire simulation dataset.

Assessing the condition of bridge infrastructure requires estimating damage-sensitive features from reliable sensor data. This study proposes to estimate natural frequencies from displacement measurements of a ground-based interferometric radar (GBR). These frequencies are determined from the damped vibration after each vehicle crossing with least squares and compared to a Frequency Domain Decomposition result. We successfully applied the approach in an exemplary measurement campaign at a bridge near Coburg (Germany) with an additional comparison to commonly used strain sensors. Since temperature greatly influences natural frequencies, linear regression is used to correct this influence. A simulation shows that GBR, combined with the least squares approach, achieves the lowest uncertainty and variation in the linear regression, indicating better damage detection results. However, the success of the damage detection highly depends on correctly determining the temperature influence, which might vary throughout the structure. Future work should further investigate the biases and variability of this influence.

This paper introduces a novel nothing-on-road (NOR) bridge weigh-in-motion (BWIM) approach with deep learning (DL) and non-invasive ground-based radar (GBR) time-series data. BWIMs allow site-specific structural health monitoring (SHM) but are usually difficult to attach and maintain. GBR measures the bridge deflection contactless. In this study, GBR and an unmanned aerial vehicle (UAV) monitor a two-span bridge in Germany to gather ground-truth data. Based on the UAV data, we determine vehicle type, lane, locus, speed, axle count, and axle spacing for single-presence vehicle crossings. Since displacement is a global response, using peak detection like conventional strain-based BWIMs is challenging. Therefore, we investigate data-driven machine learning approaches to extract the vehicle configurations directly from the displacement data. Despite a small and imbalanced real-world dataset, the proposed approaches classify, e.g., the axle count for trucks with a balanced accuracy of 76.7% satisfyingly. Additionally, we demonstrate that, for the selected bridge, high-frequency vibrations can coincide with axles crossing the junction between the street and the bridge. We evaluate whether filtering approaches via bandpass filtering or wavelet transform can be exploited for axle count and axle spacing identification. Overall, we can show that GBR is a serious contender for BWIM systems.

Natural hazards can present a significant risk to road infrastructure. This infrastructure is a fundamental component of the transportation infrastructure, with significant importance. During emergencies, society heavily relies on the functionality of the road infrastructure to facilitate evacuation and access to emergency facilities. This study introduces a versatile, multi-scale framework designed to analyze accessibility within road networks during natural hazard scenarios. The first module of the framework focuses on assessing the influence of natural hazards on road infrastructure to identify damaged or blocked road segments and intersections. It relies on near real-time information, often provided by citizen science through Volunteered Geographic Information (VGI) data and Natural Language Processing (NLP) of VGI texts. The second module conducts network analysis based on freely available Open Street Map (OSM) data, differentiating between intact and degraded road networks. Four accessibility measures are employed: betweenness centrality, closeness centrality, a free-flow assumption index, and a novel alternative routing assumption measure considering congestion scenarios. The study showcases its framework through an exemplary application in California, the United States, considering different hazard scenarios, where degraded roads and connected roads impacted by the hazard can be identified. The road extraction methodology allows the extraction of 75% to 100% of the impacted roads mentioned in VGI text messages for the respective case studies. In addition to the directly extracted impacted roads, constructing the degraded network also involves finding road segments that overlap with hazard impact zones, as these are at risk of being impacted. Conducting the network analysis with the four different measures on the intact and degraded network, changes in network accessibility due to the impacts of hazards can be identified. The results show that using each measure is justified, as each measure could demonstrate the accessibility change. However, their combination and comparison provide valuable insights. In conclusion, this study successfully addresses the challenges of developing a generic, complete framework from impact extraction to network analysis independently of the scale and characteristics of road network types.

Remote sensing techniques are frequently applied for the surveying of remote areas, where the use of conventional surveying techniques remains difficult and impracticable. In this paper, we focus on one of the remote glacier areas, namely the Tyndall Glacier area in the Southern Patagonian Icefield in Chile. Based on optical remote sensing data in the form of multi-spectral Sentinel-2 imagery, we analyze the extent of different snow and ice classes on the surface of the glacier by means of pixel-wise classification. Our study comprises three main steps: (1) Labeled Sentinel-2 compliant data are obtained from theoretical spectral reflectance curves, as there are no training data available for the investigated area; (2) Four different classification approaches are used and compared in their ability to identify the defined five snow and ice types, thereof two unsupervised approaches (k-means clustering and rule-based classification via snow and ice indices) and two supervised approaches (Linear Discriminant Analysis and Random Forest classifier); (3) We first focus on the pixel-wise classification of Sentinel-2 imagery, and we then use the best-performing approach for a multi-temporal analysis of the Tyndall Glacier area. While the achieved classification results reveal that all of the used classification approaches are suitable for detecting different snow and ice classes on the glacier surface, the multi-temporal analysis clearly reveals the seasonal development of the glacier. The change of snow and ice types on the glacier surface is evident, especially between the end of ablation season (April) and the end of accumulation season (September) in Southern Chile.

In this paper, we investigate the potential of unsupervised feature selection techniques for classification tasks, where only sparse training data are available. This is motivated by the fact that unsupervised feature selection techniques combine the advantages of standard dimensionality reduction techniques (which only rely on the given feature vectors and not on the corresponding labels) and supervised feature selection techniques (which retain a subset of the original set of features). Thus, feature selection becomes independent of the given classification task and, consequently, a subset of generally versatile features is retained. We present different techniques relying on the topology of the given sparse training data. Thereby, the topology is described with an ultrametricity index. For the latter, we take into account the Murtagh Ultrametricity Index (MUI) which is defined on the basis of triangles within the given data and the Topological Ultrametricity Index (TUI) which is defined on the basis of a specific graph structure. In a case study addressing the classification of high-dimensional hyperspectral data based on sparse training data, we demonstrate the performance of the proposed unsupervised feature selection techniques in comparison to standard dimensionality reduction and supervised feature selection techniques on four commonly used benchmark datasets. The achieved classification results reveal that involving supervised feature selection techniques leads to similar classification results as involving unsupervised feature selection techniques, while the latter perform feature selection independently from the given classification task and thus deliver generally versatile features.