Hybrid geostatistical and deep learning framework for geochemical characterization in historical mine tailings

Deep learning is a popular topic in machine learning and artificial intelligence research and has achieved remarkable results in various fields. In geological remote sensing, mineral mapping is an appealing application of hyperspectral remote sensing for geological surveyors. Whether deep learning can improve the mineral identification ability in hyperspectral remote sensing images, especially for the discrimination of spectrally similar and intimately mixed minerals, needs to be evaluated. In this study, shortwave airborne spectrographic imager (SASI) hyperspectral images of the Baiyanghe uranium deposit in Northwestern Xinjiang, China, were used as experimental data. Three deep neural network (DNN) models were designed: a fully connected neural network (FCNN), a one-dimensional convolutional neural network (1D CNN), and a one-dimensional and two-dimensional convolutional neural network (1D and 2D CNN). A sample dataset containing five minerals was constructed for model training and validation, which was divided into training, validation and test sets at a ratio of 6:2:2. The final test accuracies of the FCNN, 1D CNN, and 1D and 2D CNN were 91.24%, 93.67% and 94.77%, respectively. The three DNNs were used for mineral identification and mapping of SASI hyperspectral images of the Baiyanghe uranium mining area. The mapping results were compared with the mapping results of the support vector machine (SVM) and the mixture-tuned matched filtering (MTMF) method. Combined with the ground spectral data obtained by the spectrometer, spectral verification and interpretation were carried out on sections that the two kinds of methods identified differently. The verification results show that the mapping results of the 1D and 2D CNN were more accurate than those of the other methods. More importantly, for minerals with similar spectral characteristics, such as short-wavelength white mica and medium-wavelength white mica, the 1D and 2D CNN model had a more accurate discrimination effect than the other DNN models, indicating that the introduction of spatial information can improve the mineral identification ability in hyperspectral remote sensing images. In general, CNNs have good application prospects in geological mapping of hyperspectral remote sensing images and are worthy of further development in future work.

Bayesian optimization of one-dimensional convolutional neural networks (1D CNN) for early diagnosis of Autistic Spectrum Disorder

The availability of sufficient water supply is a challenge many municipalities have faced in recent decades and a challenge that is expected to intensify with time. While several choices remain for selecting alternatives to freshwater sources, water reclamation offers an opportunity for sustainable resource recovery. Nonetheless, tradeoffs exist in the selection of the most sustainable technology for recovering resources from wastewater when long-term impacts are taken into consideration. This article investigates the factors influencing the environmental and economic impacts of resource recovery technologies through the analysis of life cycle environmental and economic impact case studies. Key characteristics were extracted from life cycle assessment and life cycle cost case studies to evaluate the factors influencing the sustainability of the resource recovery systems. The specific design parameters include the type of resources to be recovered, technology utilized, scale of implementation, location, and end users. The design of sustainable resource recovery systems was found to be largely driven by scale, location (e.g., as it pertains to the energy mix and water quality restrictions), and the scope of the system considered. From this analysis, a decision framework for resource recovery-oriented wastewater management was developed and then applied to an existing case study to demonstrate its usability.

The sustainable management of mining wastes, a byproduct of extractive activities, represents a critical challenge in the context of the Critical Raw Materials Act (European Commission, 2023) and the transition to a circular economy. Mining waste dumps may contain significant residual amounts of ore minerals or metals, including Critical Raw Materials (CRMs), making their mapping and evaluation essential for environmental remediation and possible resource recovery. Developing detailed regional or national maps is pivotal to identifying mining waste dumps' location, typology, distribution, and spatial extent. Integrating Geographic Information System (GIS) software with complementary tools such as Google Earth, topographic maps, and orthophotos offers a comprehensive approach to efficiently identifying and analysing these sites.Sampling and characterising mining waste dumps is crucial to assessing their economic potential and environmental impact (Beltré et al., 2023). Mineralogical analyses (e.g., X-ray diffraction, Scanning Electron Microscopy, RAMAN, and Electron Microprobe) and chemical analyses (e.g., Portable X-ray Fluorescence, ICP-MS, or ICP-OES) enable the evaluation of mineral processing residues. This differentiation helps identify economically viable dumps and prioritise remediation efforts for non-viable sites with contamination risks. (Lemière et al., 2011)These methodologies are now applied to developing the Metallogenic Map of Sardinia, which will include the mapping of different mining waste dumps in Sardinia and their sampling.The crucial challenges of this project are accurately estimating dump volumes due to difficulties in identifying underlying bedrock and quantifying critical metal content. Addressing these gaps is crucial for effective resource valorisation and site rehabilitation. To date, 140 mining waste samples have been collected and are under analysis to assess their economic and environmental potential. This study integrates GIS technologies with environmental and economic assessments as a pathway to support sustainable exploitation and management of mining waste dumps, aligning with EU strategic goals for CRMs. Keywords: Critical Raw Materials, Circular Economy, Resource valorisationEuropean Commission (2023) - Study on the Critical Raw Materials for the EU. Fifth list. Final report. Rosario-Beltré, A. J., Sánchez-España, J., Rodríguez-Gómez, V., Fernández-Naranjo, F. J., Bellido-Martín, E., Adánez-Sanjuán, P., & Arranz-González, J. C. (2023). Critical Raw Materials recovery potential from Spanish mine wastes: A national-scale preliminary assessment. Journal of Cleaner Production, 407. https://doi.org/10.1016/j.jclepro.2023.137163 Lemière,, Cottard, F., & Piantone BRGM, P. (2011). Mining waste characterization in the perspective of the European mining waste directive.

Maintaining constant vigilance over arterial blood pressure (ABP) is crucial for diagnosing hypertension and other critical cardiovascular diseases. While traditional cuff-based approaches are non-invasive, they have limitations in providing continuous blood pressure monitoring. In contrast, complex ABP monitoring systems, while accurate, are primarily suitable for clinical settings due to their intrusive nature. This study introduces a groundbreaking method for generating arterial blood pressure (ABP) waveforms using remote radar signals and deep learning (DL) techniques. This approach eliminates the need for invasive procedures, wearable biosensors, and costly equipment typically associated with ABP recording. We introduce MultiResLinkNet, a segmentation model based on a one-dimensional convolutional neural network (1D CNN), specifically designed to synthesize arterial blood pressure (ABP) directly from raw radar waveforms. We trained and evaluated the end-to-end DL framework using a publicly available benchmark radar dataset containing raw radar data and corresponding physiological signals from 30 subjects across various scenarios, including Resting, Valsalva, Apnea, Tilt-up, and Tilt-down. The proposed MultiResLinkNet excelled in ABP segmentation, outperforming state-of-the-art networks in combined and individual scenarios, and produced the best average temporal and spectral correlations as well as the lowest temporal and spectral errors in nearly all scenarios’ data. Furthermore, qualitative evaluation demonstrated a strong resemblance between the synthesized and ground truth ABP waveforms. Our novel approach enables remote monitoring of critical patients continuously, especially those undergoing surgery, by predicting ABP waveforms from non-contact radar signals. This breakthrough offers significant advantages, facilitating continuous ABP monitoring without the need for invasive procedures or cumbersome wearable sensors.

Rapid on-site identification of pesticide residues in tea by one-dimensional convolutional neural network coupled with surface-enhanced Raman scattering.

Single-cell classification of foodborne pathogens using hyperspectral microscope imaging coupled with deep learning frameworks

<p>Whereas there are growing needs for mineral resources (metals for the energy and digital transitions<br>and construction materials), the mining industry must produce them from poorer, more<br>heterogeneous and more complex deposits. Therefore, volumes of mine waste produced (including<br>tailings) are also increasing and add up to waste from mining legacy. For example in Europe (x27): 732<br>Mtons of extractive waste are generated per year and more than 1.2 Btons of legacy waste are stored<br>all over the European territory. The localisation (and potential hazards) are well known and covered<br>by the inventories carried out in EU countries under the Mining Waste Directive.<br>At the same time, Europe is implementing the circular economy approach and put a lot of emphasis<br>on the resource efficiency concept. In this context, reprocessing operation to recover both metals and<br>mineral fraction is studied with the objective of combing waste management (reducing final waste<br>storage and long-term impact) and material production from secondary resources.<br>Numerous industrial experiences of reprocessing of mine waste and tailings exist all over the world to<br>recover metals such as copper, gold or critical raw materials - CRM They concern mainly active mine<br>where both primary and secondary resources are considered in profitable operations; for example in<br>Chile, South Africa, Australia. Mineral fraction recovery is often not considered which still leaves the<br>industry with a high volume of residual minerals to store and manage.<br>In addition, legacy mining waste are potentially available for reprocessing. In this case, numerous<br>mining liabilities issues need to be managed. Some of the European legacy mining waste have residual<br>valuable metals that could be recovered but some of them have very low metal contents. In Europe,<br>classical rehabilitation operations – usually at the charge of member states and local authorities – is<br>the priority and concern the reduction of instabilities and impacts to the environment including heap<br>remodelling, covering and water management with long-term treatment. Completing this risk<br>management approach by a circular economy one is a very active R&D subject in EU27.<br>This presentation will give an overview of EU research projects which tackled the legacy mining waste<br>challenge from inventory to process development. Several process flowsheets to recover metals were<br>designed and tested on several case studies with CRM – REE, Co, W, Sb, etc. Initiatives to reuse mineral<br>fraction are also underway and should be ready for commercialisation in the coming years.<br>Resources efficiency concept and the circular economy implementation starts on mining sites. In order<br>to facilitate the implementation of this approach, the technical solutions will need to be included in<br>innovative global initiatives covering also legal (liability management), environmental (Life Cycle<br>Analysis approaches) and social (acceptance) questions.</p>

We present a hardware-friendly deep learning architecture with one-dimensional convolutional neural networks (1D CNN) for fast analyzing fluorescence lifetime imaging (FLIM) data. A 1D CNN shows unparalleled advantages; they are more straightforward, quicker to train, and faster than high dimensional CNNs. 1D CNNs can be easily applied to multi-exponential fluorescence decay models. Compared with traditional least-square methods, superior performances of 1D CNNs on fluorescence lifetime image reconstruction have been validated using simulated data. We also employ the proposed 1D CNN to analyze two-photon FLIM images of functionalized gold nanoprobes in Hek293 and human prostate cancer cells. The results further demonstrate that 1D CNNs are fast and can accurately extract lifetime parameters from fluorescence signals. Our study shows that 1D CNNs have great potential in various real-time FLIM applications.

The global economy is at a transition point, moving from the traditional “make, use and discard” linear manufacturing model to a more sustainable and reusable solution that is the Circular Economy. Transitioning the electronics waste recycling industry to greater resource efficiency, re-use and circularity is championed by “closing the loop” on End-of-Life (EOL) products, recycling and re-using them as raw materials to remanufacture new products. The re-use and recycling of batteries and power packs (the lifeblood of electronic devices) from powered appliances is critical in this regard. Batteries are one of the richest sources of Critical Raw Materials (CRMs) for waste electronic recycling plants. Solutions to address this shortcoming are limited. This article proposes the RoboCRM system to address this - an automated system for battery detection which allows recyclers to close the loop on battery recovery and resource efficiency by easily identifying and sorting E-waste (Electronic waste) containing batteries from the primary waste stream. RoboCRM uses non-destructive detection methods (such as computer vision systems) in conjunction with pattern recognition and an artificial intelligence engine to achieve this. Once identified and categorised, these battery-powered appliances can be processed to support greater recovery of raw materials and CRMs. This will help close the loop on this aspect of the E-waste stream. As a result, the E-waste recycling industry will be able to integrate circular economy principles, resource recovery and re-use into their existing models in a seamless way, creating new jobs in the industry and producing a highly skilled circular economy workforce. The RoboCRM integrates computer vision and imaging solutions with robotics, artificial intelligence and machine learning technologies, in order to create a breakthrough system which will become indispensable for the global recycling industry.

This report is a background document used by several European Commission services to prepare the EC report on critical raw materials and the circular economy, a commitment of the European Commission made in its Communication ‘EU action plan for the Circular Economy'. It represents a JRC contribution to the Raw Material Initiative and to the EU Circular Economy Action Plan. It combines the results of several research programmes and activities of the JRC on critical raw materials in a context of circular economy, for which a large team has contributed in terms of data and knowledge developments. Circular use of critical raw materials in the EU is analysed, also taking a sectorial perspective. The following sectors are analysed in more detail: extractive waste, landfills, electric and electronic equipment, batteries, automotive, renewable energy, defence and chemicals and fertilisers. Conclusions and opportunities for further work are also presented.

Non-destructive estimation for Kyoho grape shelf-life using Vis/NIR hyperspectral imaging and deep learning algorithm

Automatic identification of abnormal and irregular heart rhythms is necessary to reduce mortality. Tachyarrhythmia is a type of abnormally fast heartbeat that can be detected using electrocardiogram (ECG) signals. In the elderly, life-threatening tachyarrhythmia such as ventricular fibrillation (VFIB), atrial fibrillation (AFIB), and atrial flutter (AFL) can lead to sudden cardiac arrest. Here, we present a hybrid deep learning (HDL) model for automatic identification of tachyarrhythmia rhythms from heart rate variability (HRV) datasets based on a one-dimensional convolution neural network (1D CNN) and a long-term short-term memory (LSTM) model. In this study, we used the HRV database with five-second windows as input data for our HDL model. Four different statistical parameters have been used to determine the model efficiency: The average accuracy is 99.19%, the average precision is 91.75%, the recall is 93.63%, and the F1 score is 92.71%. The overall accuracy of the experiment was 98.4%. This model outperformed other state-of-the-art models. As a result, this method can be useful in clinical systems of cardiological care.KeywordsAFIBAFLVFIBHRVCNNLSTM

RUL prediction of lithium ion battery based on CEEMDAN-CNN BiLSTM model

The application of laser-induced fluorescence (LIF) combined with machine learning methods can make up for the shortcomings of traditional hydrochemical methods in the accurate and rapid identification of mine water inrush in coal mines. However, almost all of these methods require preprocessing such as principal component analysis (PCA) or drawing the spectral map as an essential step. Here, we provide our solution for the classification of mine water inrush, in which a one-dimensional convolutional neural network (1D CNN) is trained to automatically identify mine water inrush according to the LIF spectroscopy without the need for preprocessing. First, the architecture and parameters of the model were optimized and the 1D CNN model containing two convolutional blocks was determined to be the best model for the identification of mine water inrush. Then, we evaluated the performance of the 1D CNN model using the LIF spectral dataset of mine water inrush containing 540 training samples and 135 test samples, and we found that all 675 samples could be accurately identified. Finally, superior classification performance was demonstrated by comparing with a traditional machine learning algorithm (genetic algorithm-support vector machine) and a deep learning algorithm (two-dimensional convolutional neural network). The results show that LIF spectroscopy combined with 1D CNN can be used for the fast and accurate identification of mine water inrush without the need for complex pretreatments.