Helmholtz Institute Freiberg for Resource Technology

Hyperspectral image (HSI) classification has become a hot topic in the field of remote sensing. In general, the complex characteristics of hyperspectral data make the accurate classification of such data challenging for traditional machine learning methods. In addition, hyperspectral imaging often deals with an inherently nonlinear relation between the captured spectral information and the corresponding materials. In recent years, deep learning has been recognized as a powerful feature-extraction tool to effectively address nonlinear problems and widely used in a number of image processing tasks. Motivated by those successful applications, deep learning has also been introduced to classify HSIs and demonstrated good performance. This survey paper presents a systematic review of deep learning-based HSI classification literatures and compares several strategies for this topic. Specifically, we first summarize the main challenges of HSI classification which cannot be effectively overcome by traditional machine learning methods, and also introduce the advantages of deep learning to handle these problems. Then, we build a framework that divides the corresponding works into spectral-feature networks, spatial-feature networks, and spectral-spatial-feature networks to systematically review the recent achievements in deep learning-based HSI classification. In addition, considering the fact that available training samples in the remote sensing field are usually very limited and training deep networks require a large number of samples, we include some strategies to improve classification performance, which can provide some guidelines for future studies on this topic. Finally, several representative deep learning-based classification methods are conducted on real HSIs in our experiments.

Several machine-learning algorithms have been proposed for remote sensing image classification during the past two decades. Among these machine learning algorithms, Random Forest (RF) and Support Vector Machines (SVM) have drawn attention to image classification in several remote sensing applications. This article reviews RF and SVM concepts relevant to remote sensing image classification and applies a meta-analysis of 251 peer-reviewed journal papers. A database with more than 40 quantitative and qualitative fields was constructed from these reviewed papers. The meta-analysis mainly focuses on 1) the analysis regarding the general characteristics of the studies, such as geographical distribution, frequency of the papers considering time, journals, application domains, and remote sensing software packages used in the case studies, and 2) a comparative analysis regarding the performances of RF and SVM classification against various parameters, such as data type, RS applications, spatial resolution, and the number of extracted features in the feature engineering step. The challenges, recommendations, and potential directions for future research are also discussed in detail. Moreover, a summary of the results is provided to aid researchers to customize their efforts in order to achieve the most accurate results based on their thematic applications.

The foundation model has recently garnered significant attention due to its potential to revolutionize the field of visual representation learning in a self-supervised manner. While most foundation models are tailored to effectively process RGB images for various visual tasks, there is a noticeable gap in research focused on spectral data, which offers valuable information for scene understanding, especially in remote sensing (RS) applications. To fill this gap, we created for the first time a universal RS foundation model, named SpectralGPT, which is purpose-built to handle spectral RS images using a novel 3D generative pretrained transformer (GPT). Compared to existing foundation models, SpectralGPT 1) accommodates input images with varying sizes, resolutions, time series, and regions in a progressive training fashion, enabling full utilization of extensive RS Big Data; 2) leverages 3D token generation for spatial-spectral coupling; 3) captures spectrally sequential patterns via multi-target reconstruction; and 4) trains on one million spectral RS images, yielding models with over 600 million parameters. Our evaluation highlights significant performance improvements with pretrained SpectralGPT models, signifying substantial potential in advancing spectral RS Big Data applications within the field of geoscience across four downstream tasks: single/multi-label scene classification, semantic segmentation, and change detection.

Lithium-ion batteries (LIBs) are currently one of the most important electrochemical energy storage devices, powering electronic mobile devices and electric vehicles alike. However, there is a remarkable difference between their rate of production and rate of recycling. At the end of their lifecycle, only a limited number of LIBs undergo any recycling treatment, with the majority go to landfills or being hoarded in households. Further losses of LIB components occur because the the state-of-the-art LIB recycling processes are limited to components with high economic value, e.g., Co, Cu, Fe, and Al. With the increasing popularity of concepts such as “circular economy” (CE), new LIB recycling systems have been proposed that target a wider spectrum of compounds, thus reducing the environmental impact associated with LIB production. This review work presents a discussion of the current practices and some of the most promising emerging technologies for recycling LIBs. While other authoritative reviews have focused on the description of recycling processes, the aim of the present was is to offer an analysis of recycling technologies from a CE perspective. Consequently, the discussion is based on the ability of each technology to recover every component in LIBs. The gathered data depicted a direct relationship between process complexity and the variety and usability of the recovered fractions. Indeed, only processes employing a combination of mechanical processing, and hydro- and pyrometallurgical steps seemed able to obtain materials suitable for LIB (re)manufacture. On the other hand, processes relying on pyrometallurgical steps are robust, but only capable of recovering metallic components.

By considering the spectral signature as a sequence, recurrent neural networks (RNNs) have been successfully used to learn discriminative features from hyperspectral images (HSIs) recently. However, most of these models only input the whole spectral bands into RNNs directly, which may not fully explore the specific properties of HSIs. In this paper, we propose a cascaded RNN model using gated recurrent units to explore the redundant and complementary information of HSIs. It mainly consists of two RNN layers. The first RNN layer is used to eliminate redundant information between adjacent spectral bands, while the second RNN layer aims to learn the complementary information from nonadjacent spectral bands. To improve the discriminative ability of the learned features, we design two strategies for the proposed model. Besides, considering the rich spatial information contained in HSIs, we further extend the proposed model to its spectral-spatial counterpart by incorporating some convolutional layers. To test the effectiveness of our proposed models, we conduct experiments on two widely used HSIs. The experimental results show that our proposed models can achieve better results than the compared models.

Hyperspectral images (HSIs) provide detailed spectral information through hundreds of (narrow) spectral channels (also known as dimensionality or bands), which can be used to accurately classify diverse materials of interest. The increased dimensionality of such data makes it possible to significantly improve data information content but provides a challenge to conventional techniques (the so-called curse of dimensionality) for accurate analysis of HSIs.

Vision transformers (ViTs) have been trending in image classification tasks due to their promising performance when compared to convolutional neural networks (CNNs). As a result, many researchers have tried to incorporate ViTs in hyperspectral image (HSI) classification tasks. To achieve satisfactory performance, close to that of CNNs, transformers need fewer parameters. ViTs and other similar transformers use an external classification (CLS) token which is randomly initialized and often fails to generalize well, whereas other sources of multimodal datasets, such as light detection and ranging (LiDAR) offer the potential to improve these models by means of a CLS. In this paper, we introduce a new multimodal fusion transformer (MFT) network which comprises a multihead cross patch attention (mCrossPA) for HSI land-cover classification. Our mCrossPA utilizes other sources of complementary information in addition to the HSI in the transformer encoder to achieve better generalization. The concept of tokenization is used to generate CLS and HSI patch tokens, helping to learn a distinctive representation in a reduced and hierarchical feature space. Extensive experiments are carried out on widely used benchmark datasets i.e., the University of Houston, Trento, University of Southern Mississippi Gulfpark (MUUFL), and Augsburg. We compare the results of the proposed MFT model with other state-of-the-art transformers, classical CNNs, and conventional classifiers models. The superior performance achieved by the proposed model is due to the use of multihead cross patch attention. The source code will be made available publicly at https://github.com/AnkurDeria/MFT.

In this article, we propose an efficient and effective framework to fuse hyperspectral and light detection and ranging (LiDAR) data using two coupled convolutional neural networks (CNNs). One CNN is designed to learn spectral-spatial features from hyperspectral data, and the other one is used to capture the elevation information from LiDAR data. Both of them consist of three convolutional layers, and the last two convolutional layers are coupled together via a parameter-sharing strategy. In the fusion phase, feature-level and decision-level fusion methods are simultaneously used to integrate these heterogeneous features sufficiently. For the feature-level fusion, three different fusion strategies are evaluated, including the concatenation strategy, the maximization strategy, and the summation strategy. For the decision-level fusion, a weighted summation strategy is adopted, where the weights are determined by the classification accuracy of each output. The proposed model is evaluated on an urban data set acquired over Houston, USA, and a rural one captured over Trento, Italy. On the Houston data, our model can achieve a new record overall accuracy (OA) of 96.03%. On the Trento data, it achieves an OA of 99.12%. These results sufficiently certify the effectiveness of our proposed model.

Convolutional neural networks (CNNs) have been widely used for hyperspectral image classification. As a common process, small cubes are first cropped from the hyperspectral image and then fed into CNNs to extract spectral and spatial features. It is well known that different spectral bands and spatial positions in the cubes have different discriminative abilities. If fully explored, this prior information will help improve the learning capacity of CNNs. Along this direction, we propose an attention-aided CNN model for spectral-spatial classification of hyperspectral images. Specifically, a spectral attention subnetwork and a spatial attention subnetwork are proposed for spectral and spatial classifications, respectively. Both of them are based on the traditional CNN model and incorporate attention modules to aid networks that focus on more discriminative channels or positions. In the final classification phase, the spectral classification result and the spatial classification result are combined together via an adaptively weighted summation method. To evaluate the effectiveness of the proposed model, we conduct experiments on three standard hyperspectral data sets. The experimental results show that the proposed model can achieve superior performance compared with several state-of-the-art CNN-related models.

In recent years, airborne and spaceborne hyperspectral imaging systems have advanced in terms of spectral and spatial resolution, which makes the data sets they produce a valuable source for land cover classification. The availability of hyperspectral data with fine spatial resolution has revolutionized hyperspectral image (HSI) classification techniques by taking advantage of both spectral and spatial information in a single classification framework.

So far, a large number of advanced techniques have been developed to enhance and extract the spatially semantic information in hyperspectral image processing and analysis. However, locally semantic change, such as scene composition, relative position between objects, spectral variability caused by illumination, atmospheric effects, and material mixture, has been less frequently investigated in modeling spatial information. Consequently, identifying the same materials from spatially different scenes or positions can be difficult. In this article, we propose a solution to address this issue by locally extracting invariant features from hyperspectral imagery (HSI) in both spatial and frequency domains, using a method called invariant attribute profiles (IAPs). IAPs extract the spatial invariant features by exploiting isotropic filter banks or convolutional kernels on HSI and spatial aggregation techniques (e.g., superpixel segmentation) in the Cartesian coordinate system. Furthermore, they model invariant behaviors (e.g., shift, rotation) by the means of a continuous histogram of oriented gradients constructed in a Fourier polar coordinate. This yields a combinatorial representation of spatial-frequency invariant features with application to HSI classification. Extensive experiments conducted on three promising hyperspectral data sets (Houston2013 and Houston2018) to demonstrate the superiority and effectiveness of the proposed IAP method in comparison with several state-of-the-art profile-related techniques. The codes will be available from the website: https://sites.google.com/view/danfeng-hong/data-code.

Several epidemiological models are being used around the world to project the number of infected individuals and the mortality rates of the COVID-19 outbreak. Advancing accurate prediction models is of utmost importance to take proper actions. Due to the lack of essential data and uncertainty, the epidemiological models have been challenged regarding the delivery of higher accuracy for long-term prediction. As an alternative to the susceptible-infected-resistant (SIR)-based models, this study proposes a hybrid machine learning approach to predict the COVID-19, and we exemplify its potential using data from Hungary. The hybrid machine learning methods of adaptive network-based fuzzy inference system (ANFIS) and multi-layered perceptron-imperialist competitive algorithm (MLP-ICA) are proposed to predict time series of infected individuals and mortality rate. The models predict that by late May, the outbreak and the total morality will drop substantially. The validation is performed for 9 days with promising results, which confirms the model accuracy. It is expected that the model maintains its accuracy as long as no significant interruption occurs. This paper provides an initial benchmarking to demonstrate the potential of machine learning for future research.

Circular economy's (CE) noble aims maximize resource efficiency (RE) by, for example, extending product life cycles and using wastes as resources. Modern society's vast and increasing amounts of waste and consumer goods, their complexity, and functional material combinations are challenging the viability of the CE despite various alternative business models promising otherwise. The metallurgical processing of CE-enabling technologies requires a sophisticated and agile metallurgical infrastructure. The challenges of reaching a CE are highlighted in terms of, e.g., thermodynamics, transfer processes, technology platforms, digitalization of the processes of the CE stakeholders, and design for recycling (DfR) based on a product (mineral)-centric approach, highlighting the limitations of material-centric considerations. Integrating product-centric considerations into the water, energy, transport, heavy industry, and other smart grid systems will maximize the RE of future smart sustainable cities, providing the fundamental detail for realizing and innovating the United Nation's Sustainability Development Goals.

Froth flotation of scheelite has regained new focus since the 2010s and research regarding floatability and reagents has made great progress over the years. The main objective was and remains the selective flotation of scheelite from other calcium-bearing minerals, in particular calcite, fluorite and apatite. Due to similar properties, most attempts have limited success or only specific application (linked to a type of ore or a location). This article aims at reviewing all general physical-chemical information on froth flotation of scheelite, including electrokinetic properties, influence of pH and already existing reagents as well as ones still under examination. It appears that chelating or mixed collectors and modified versions of sodium silicate and quebracho hold great promise for scheelite flotation, while the use of said depressants and/or promoters seems inevitable.

Graphite as the most common polymorph of naturally occurring crystalline carbon is required for many different applications such as batteries, refractories, electrical products, and pencils. Graphite resources are currently being subjected to intensive exploration to help meet rapidly growing global demand – and graphite has made it onto the list of critical raw materials as issued by the European Union. Graphite ore is mostly beneficiated using flotation separation techniques. The increasing demand for high-grade graphite products with up to 99.99% carbon has resulted in the development of various approaches to remove impurities even to parts per million range. This paper considers separation and purification techniques that are currently employed for graphite mineral beneficiation, and identifies areas in need of further research.

The exposome encompasses an individual's exposure to exogenous chemicals, as well as endogenous chemicals that are produced or altered in response to external stressors. While the exposome concept has been established for human health, its principles can be extended to include broader ecological issues. The assessment of exposure is tightly interlinked with hazard assessment. Here, we explore if mechanistic understanding of the causal links between exposure and adverse effects on human health and the environment can be improved by integrating the exposome approach with the adverse outcome pathway (AOP) concept that structures and organizes the sequence of biological events from an initial molecular interaction of a chemical with a biological target to an adverse outcome. Complementing exposome research with the AOP concept may facilitate a mechanistic understanding of stress-induced adverse effects, examine the relative contributions from various components of the exposome, determine the primary risk drivers in complex mixtures, and promote an integrative assessment of chemical risks for both human and environmental health.

The classification accuracy of remote sensing data can be increased by integrating ancillary data provided by multisource acquisition of the same scene. We propose to merge the spectral and spatial content of hyperspectral images (HSIs) with elevation information from light detection and ranging (LiDAR) measurements. In this paper, we propose to fuse the data sets using orthogonal total variation component analysis (OTVCA). Extinction profiles are used to automatically extract spatial and elevation information from HSI and rasterized LiDAR features. The extracted spatial and elevation information is then fused with spectral information using the OTVCA-based feature fusion method to produce the final classification map. The extracted features have high dimension, and therefore OTVCA estimates the fused features in a lower dimensional space. OTVCA also promotes piece-wise smoothness while maintaining the spatial structures. Both attributes are important to provide homogeneous regions in the final classification maps. We benchmark the proposed approach (OTVCA-fusion) with an urban data set captured over an urban area in Houston/USA and a rural region acquired in Trento/Italy. In the experiments, OTVCA-fusion is evaluated using random forest and support vector machine classifiers. Our experiments demonstrate the ability of OTVCA-fusion to produce accurate classification maps while using fewer features compared with other approaches investigated in this paper.

Drone-borne hyperspectral imaging is a new and promising technique for fast and precise acquisition, as well as delivery of high-resolution hyperspectral data to a large variety of end-users. Drones can overcome the scale gap between field and air-borne remote sensing, thus providing high-resolution and multi-temporal data. They are easy to use, flexible and deliver data within cm-scale resolution. So far, however, drone-borne imagery has prominently and successfully been almost solely used in precision agriculture and photogrammetry. Drone technology currently mainly relies on structure-from-motion photogrammetry, aerial photography and agricultural monitoring. Recently, a few hyperspectral sensors became available for drones, but complex geometric and radiometric effects complicate their use for geology-related studies. Using two examples, we first show that precise corrections are required for any geological mapping. We then present a processing toolbox for frame-based hyperspectral imaging systems adapted for the complex correction of drone-borne hyperspectral imagery. The toolbox performs sensor- and platform-specific geometric distortion corrections. Furthermore, a topographic correction step is implemented to correct for rough terrain surfaces. We recommend the c-factor-algorithm for geological applications. To our knowledge, we demonstrate for the first time the applicability of the corrected dataset for lithological mapping and mineral exploration.

Abstract Recent landslide detection studies have focused on pixel-based deep learning (DL) approaches. In contrast, intuitive annotation of landslides from satellite imagery is based on distinct features rather than individual pixels. This study examines the feasibility of the integration framework of a DL model with rule-based object-based image analysis (OBIA) to detect landslides. First, we designed a ResU-Net model and then trained and tested it in the Sentinel-2 imagery. Then we developed a simple rule-based OBIA with only four rulesets, applying it first to the original image dataset and then to the same dataset plus the resulting ResU-Net heatmap. The value of each pixel in the heatmap refers to the probability that the pixel belongs to either landslide or non-landslide classes. Thus, we evaluate three scenarios: ResU-Net, OBIA, and ResU-Net-OBIA. The landslide detection maps from three different classification scenarios were compared against a manual landslide inventory map using thematic accuracy assessment metrics: precision, recall, and f1-score. Our experiments in the testing area showed that the proposed integration framework yields f1-score values 8 and 22 percentage points higher than those of the ResU-Net and OBIA approaches, respectively.

Decarbonization of economy is intended to reduce the consumption of non-renewable energy sources and emissions from them. One of the major components of decarbonization are “green energy” technologies, e.g. wind turbines and electric vehicles. However, they themselves create new sustainability challenges, e.g. use of green energy contributes to the reduction of consumption of fossil fuels, on one hand, but at the same time it increases demand for permanent magnets containing considerable amounts of rare earth elements (REEs). This article provides the first global analysis of environmental impact of using rare earth elements in green energy technologies. The analysis was performed applying system dynamics modelling methodology integrated with life cycle assessment and geometallurgical approach. We provide evidence that an increase by 1% of green energy production causes a depletion of REEs reserves by 0.18% and increases GHG emissions in the exploitation phase by 0.90%. Our results demonstrate that between 2010 and 2020, the use of permanent magnets has resulted cumulatively in 32 billion tonnes CO2-equivalent of GHG emissions globally. It shows that new approaches to decarbonization are still needed, in order to ensure sustainability of the process. The finding highlights a need to design and implement various measures intended to increase REEs reuse, recycling (currently below 1%), limit their dematerialization, increase substitution and develop new elimination technologies. Such measures would support the development of appropriate strategies for decarbonization and environmentally sustainable development of green energy technologies.