Search Results (763)

In space environments, microgravity, high radiation, and weak magnetic fields induce behavioral alterations in animals, resulting in erratic movement patterns that complicate tracking. These challenges impede accurate behavioral analysis, especially in multi-object scenarios. To address this issue, this study proposes a deep learning-based multi-object tracking (MOT) framework specifically designed for space animals. The proposed method decouples appearance and motion features through dual-stream inputs and employs modality-specific encoders (MSEs), which are fused via a heterogeneous graph network to model cross-modal spatio-temporal relationships. Additionally, an object re-detection module is integrated to maintain identity continuity during occlusions or rapid movements. This approach is validated using public datasets of space-observed Drosophila and zebrafish, with experimental results demonstrating superior performance compared with existing tracking methods. This work highlights the potential of artificial intelligence as a valuable tool in behavioral studies, enabling reliable animal tracking and analysis under extreme space conditions and supporting future research in space life sciences. Full article

For the pulsed data streams emitted by multiple signal sources that generate aliasing, traditional density clustering algorithms have the problems of poor clustering effect, heavy reliance on manual experience to set the parameters, and the need to carry out density clustering every time new data are input, resulting in a huge amount of computation. Therefore, an online density clustering algorithm based on the improved golden sine whale optimization is proposed. First, by adding new parameters to the density clustering algorithm, the neighborhood is changed from a single parameter Eps to a joint decision of the parameters Eps and $\theta$, which avoids cross-cluster expansion by more flexibly delimiting the neighborhood range. The improved golden sine whale optimization algorithm is then used to obtain the optimal parameter solution of the DBSCAN algorithm. Finally, the idea of flow clustering is introduced to determine whether a pulse belongs to a valid library, an outlier library, or an inactive library by comparing the distance between the input pulse and each cluster center, effectively reducing the number of pulses required for analysis. The experiment proves that the algorithm improves the sorting accuracy by 57.7% compared to the DBSCAN algorithm and 37.8% compared to the WOA-DBSCAN algorithm. Full article

Background: Studies have shown that humans can rapidly learn the shape of new objects or adjust their behavior when encountering novel situations. Research on visual cognition in the brain further indicates that the ventral visual pathway plays a critical role in core object recognition. While existing studies often focus on microscopic simulations of individual neural structures, few adopt a holistic, system-level perspective, making it difficult to achieve robust few-shot learning capabilities. Method: Inspired by the mechanisms and processes of the ventral visual stream, this paper proposes a computational model with a macroscopic neural architecture for few-shot learning. We reproduce the feature extraction functions of V1 and V2 using a well-trained Vision Transformer (ViT) and model the neuronal activity in V4 and IT using two neural fields. By connecting these neurons based on Hebbian learning rules, the proposed model stores the feature and category information of the input samples during support training. Results: By employing a scale adaptation strategy, the proposed model emulates visual neural mechanisms, enables efficient learning, and outperforms state-of-the-art few-shot learning algorithms in comparative experiments on real-world image datasets, demonstrating human-like learning capabilities. Conclusion: Experimental results demonstrate that our ventral-stream-inspired machine-learning model achieves effective few-shot learning on real-world datasets. Full article

Nowadays, guided acoustic waves (GAW) are used for many sensor and actuator applications. The use of numerical methods can facilitate the development and optimization process enormously. In this work, a universally applicable finite element method (FEM)-based model is introduced to determine the dispersion relations of guided acoustic waves. A 2-dimensional unit cell model with Floquet periodicity is used to calculate the corresponding band structure diagrams. Starting from a free plate the model is expanded to encompass single-sided fluid loading. Followed by a straightforward algorithm for post-processing, the data is presented. Additionally, a parametric optimizer is used to adapt the simulations to experimental data measured by a laser Doppler vibrometer on an aluminum plate. Finally, the accuracy of the FEM model is compared to two reference models, achieving good consistency. In the case of the fluid-loaded model, the behavior of critical interactions between the dispersion curves and model-based artifacts is discussed. This approach can be used to model 2D structures like phononic crystals, which cannot be simulated by common GAW models. Moreover, this method can be used as input for advanced multiphysics simulations, including acoustic streaming applications. Full article

Blackwater rivers are enriched in humic acids and impoverished in nutrients, sometimes discharging into oceans. Brazil has a coastal zone of about 8700 km, with several blackwater rivers discharging into the Atlantic Ocean, in addition to the Rio Negro of the northern Amazon basin, which is the largest (about 1700 km long) and best-known tropical backwater river. On the other hand, only a few attempts have been made to deal with their hydrochemical composition and how it is related to the hydrochemistry of different water bodies nearby. This paper focuses on a sector of the Atlantic Ocean shore occurring in São Paulo State, enclosing two important Ecological Reserves, i.e., the Restinga State Park of Bertioga and the State Park of Serra do Mar–São Sebastião Nucleus, located at Bertioga and São Sebastião cities, respectively. Physicochemical parameters such as pH and electrical conductivity, as well as the composition of major constituents like sodium, potassium, calcium, magnesium, bicarbonate, chloride, sulfate, nitrate, etc., have been evaluated in two blackwater rivers and one blackwater stream to compare their relative inputs into the Atlantic Ocean. Traditional hydrogeochemical diagrams such as the Piper, Schoeller, Gibbs, van Wirdum, and Wilcox graphs were utilized for investigating the major features of the blackwater’s composition, revealing in some cases that they suffer an accentuated influence of the constituents occurring in the Atlantic Ocean waters, due to backward currents (coastal upwelling or tidal currents). Another highlight of this paper is the measurement of an enhanced concentration of dissolved iron in one blackwater sample analyzed, reaching a value of 1.9 mg/L. Such a finding has also been often reported in the literature for blackwater rivers and streams, as humic and fulvic acids are used to bind Fe³⁺, keeping it in solution. Nowadays, iron in solution has been considered a very important element acting as a natural fertilizer of the coastal ocean because it is an essential nutrient to marine phytoplankton. Full article

The construction sector, facing a persistent productivity gap compared to other industries, is hindered by fragmented value streams, inconsistent performance metrics, and the limited scalability of process improvements. We introduce a pioneering, four-tiered hierarchical productivity framework to respond to these challenges. This innovative approach integrates operational, tactical, strategic, and normative layers. At its core, the framework applies standardised, repeatable process steps—mapped using Value Stream Mapping (VSM)—to capture key indicators such as input efficiency, output effectiveness, and First-Time Quality (FTQ). These are then aggregated through takt time compliance, schedule reliability, and workload balance to evaluate trade synchronisation and flow stability. Higher-level metrics—flow efficiency, multi-resource utilisation, and ESG-linked performance—are integrated into an Overall Productivity Index (OPI). Building on a modular production model, the proposed framework supports real-time sensing, AI-driven monitoring, and intelligent process control, as demonstrated through an empirical case study of continuous process monitoring for Kelly drilling operations. This validation illustrates how sensor-equipped machinery and machine learning algorithms can automate data capture, map observed activities to standardised process steps, and detect productivity deviations in situ. This paper contributes to a multi-scalar measurement architecture that links micro-level execution with macro-level decision-making. It provides a foundation for real-time monitoring, performance-based coordination, and data-driven innovation. The framework is applicable across modular construction, digital twins, and platform-based delivery models, offering benefits beyond specialised foundation work to all construction trades. Grounded in over a century of productivity research, the approach demonstrates how emerging technologies can deliver measurable and scalable improvements. Framing productivity as an integrative, actionable metric enables sector-wide performance gains. The framework supports construction firms, technology providers, and policymakers in advancing robust, outcome-oriented innovation strategies. Full article

Mangrove forests are among the world’s most vital coastal ecosystems. Mapping mangrove cover from local to global scales using spectral data and machine learning models is a well-established method. While non-spectral contextual datasets (spatial features) have also been incorporated into such models, the contribution of these additional features to improving mangrove mapping remains underexplored. Using the Mesoamerican Reef Region as a case study, we evaluate the effectiveness of incorporating spatial features in binary mangrove classification to enhance mapping accuracy. We compared an aspatial model that includes only spectral data with three spatial models: two included features such as geographic coordinates, elevation, and proximity to coastlines and streams, while the third integrated a geostatistical approach using Inverse Distance Weighted (IDW) interpolation. Spectral inputs included bands and indices derived from Sentinel-1 and Sentinel-2, and all models were implemented using the Random Forest algorithm in Google Earth Engine. Results show that spatial features reduced omission errors without increasing commission errors, enhancing the model’s ability to capture spatial variability. Models using geographic coordinates and elevation performed comparably to those with additional environmental variables, with storm frequency and distance to streams emerging as important predictors in the Mesoamerican Reef region. In contrast, the IDW-based model underperformed, likely due to overfitting and limited representation of local spectral variation. Spatial analyses show that models incorporating spatial features produced more continuous mangrove patches and removed some false positives in non-mangrove areas. These findings highlight the value of spatial features in improving classification accuracy, especially in regions with ecologically diverse mangroves across varied environments. By integrating spatial context, these models support more accurate, locally relevant mangrove maps that are essential for effective conservation and management. Full article

The information on the seasonal variability of the chemical composition of the Arctic rivers is necessary for the proper assessment of the status of river runoff and the influence of anthropogenic and natural factors. Spring freshet is an especially important period for the Arctic rivers with a sharp maximum of water discharge. The Kolyma River is the least studied large river with a basin located solely in the permafrost zone. The change in the concentration of dissolved organic carbon (DOC), major, trace, and rare earth (RE) elements was studied at the peak and waning of the spring freshet of 2024 in the lower reaches of the Kolyma River. The concentration of elements was determined in filtrates <0.45 μm and in suspended solids > 0.45 μm. The content of coarse colloids (0.05–0.45 μm) was estimated by the intensity of dynamic light scattering (DLS). It was shown that the freshet peak is characterized by a minimal specific conductivity, concentration of major cations, and chemical elements migrating mainly in solution (Li, Sr, and Ba). During the freshet decline, the concentration of these elements increases with dynamics depending on the water exchange. The waters from the Kolyma River main stream have a maximal content of coarse colloids and concentration of <0.45 μm forms of hydrolysates (Al, Ti, Fe, Mn, REEs, Zr, Y, Sc, and Th), DOC, P, and heavy metals (Cu, Ni, Cd, and Co) at the freshet peak. A decrease of 8–10 times for hydrolysates and coarse colloids (0.05–0.45 μm) and of 3–6 times for heavy metals was observed at the freshet waning during the first half of June. This indicates a large-scale accumulation of easy soluble forms of hydrolysates, DOC, and heavy metals in the seasonal thawing topsoil layer on the catchment upstream in the previous summer, with a flush out of these elements at the freshet peak of the current year. In the large floodplain watercourse Panteleikha River, the change in concentration of major cations and REEs, Zr, Y, Sc, and Th at the freshet is less accented compared with the Kolyma River main stream due to a slower water exchange. Yet, <0.45 μm forms of Fe, Mn, Co, As, V, and P show an increase of 4–6 times in the Panteleikha River in the second half of June compared with the freshet peak, which indicates an additional input of these elements from the thawing floodplain landscapes and bottom sediments of floodplain watercourses. The concentration of the majority of chemical elements in suspended matter (>0.45 μm) of the Kolyma River is rather stable during the high-water period. The relative stability in the chemical composition of the suspended solids means that the content of the suspension and not its composition is the key to the share of dissolved and suspended forms of chemical elements in the Kolyma River runoff. Full article

This study aims to evaluate eco-innovative solutions in the fertilizer industry that allow for waste valorization in the context of a resource-efficient circular economy. A comprehensive reuse strategy was developed for low-rank peat and post-cultivation horticultural mineral wool, involving the extraction of valuable humic substances from peat and residual nutrients from used mineral wool, followed by the use of both post-extraction residues to produce organic–mineral substrates. The resulting products/semifinished products were characterized in terms of their composition and properties, which met the requirements necessary to obtain the admission of this type of product to the market in accordance with the Regulation of the Minister for Agriculture and Rural Development of 18 June 2008 on the implementation of certain provisions of the Act on fertilizers and fertilization (Journal of Laws No 119, item 765). Elemental analysis, FTIR spectroscopy, and solid-state CP-MAS ¹³C NMR spectroscopy suggest that post-extraction peat has a relatively condensed structure with a high C content (47.4%) and a reduced O/C atomic ratio and is rich in alkyl-like matter (63.2%) but devoid of some functional groups in favor of extracted fulvic acids. Therefore, it remains a valuable organic biowaste, which, in combination with post-extraction waste mineral wool in a ratio of 60:40 and possibly the addition of mineral nutrients, allows us to obtain a completely new substrate with a bulk density of 264 g/m³, a salinity of 7.8 g/dm³ and a pH of 5.3, with an appropriate content of heavy metals and with no impurities, meeting the requirements of this type of product. A liquid fertilizer based on an extract containing previously recovered nutrients also meets the criteria in terms of quality and content of impurities and can potentially be used as a fertilizing product suitable for agricultural crops. This study demonstrates a feasible pathway for transforming specific waste streams into valuable agricultural inputs, contributing to environmental protection and sustainable production. The production of a new liquid fertilizer using nutrients recovered from post-cultivation mineral wool and the preparation of an organic–mineral substrate using post-extraction solid residue is a rational strategy for recycling hard-to-biodegrade end-of-life products. Full article

Understanding groundwater and surface water interaction is critical for managing water resources, particularly in water-stressed and rapidly urbanizing areas, such as many parts of Africa. A survey was conducted of borehole, spring, seep and river water radon, δ²H, δ¹⁸O and field parameters in the Jukskei River catchment, Johannesburg. Average values of electrical conductivity (EC) were 274 and 411 μS·cm⁻¹ for groundwater and surface water, and similarly for radon, 37,000 and 1100 Bq·m⁻³, with a groundwater high of 196,000 Bq·m⁻³ associated with a structural lineament. High radon was a good indicator of baseflow, highest at the end of the rainy season (March) and lowest at the end of the dry season (September), with the FINIFLUX model computing groundwater inflow as 2.5–4.7 L·m⁻¹s⁻¹. High EC was a poorer indicator of baseflow, also considering the possibility of wastewater with high EC, typical in urban areas. Groundwater δ²H and δ¹⁸O values are spread widely, suggesting recharge from both normal and unusual rainfall periods. A slight shift from the local meteoric water line indicates light evaporation during recharge. Surface water δ²H and δ¹⁸O is clustered, pointing to regular groundwater input along the stream, supporting the findings from radon. Given the importance of groundwater, further study using the same parameters or additional analytes is advisable in the urban area of Johannesburg or other cities. Full article

Recently, deep learning algorithms have been increasingly applied in construction for activity recognition, particularly for excavators, to automate processes and enhance safety and productivity through continuous monitoring of earthmoving activities. These deep learning algorithms analyze construction videos to classify excavator activities for earthmoving purposes. However, previous studies have solely focused on single-source external videos, which limits the activity recognition capabilities of the deep learning algorithm. This paper introduces a novel multi-modal deep learning-based methodology for recognizing excavator activities, utilizing multi-stream input data. It processes point clouds and RGB images using the two-stream long short-term memory convolutional neural network (CNN-LSTM) method to extract spatiotemporal features, enabling the recognition of excavator activities. A comprehensive dataset comprising 495,000 video frames of synchronized RGB and point cloud data was collected across multiple construction sites under varying conditions. The dataset encompasses five key excavator activities: Approach, Digging, Dumping, Idle, and Leveling. To assess the effectiveness of the proposed method, the performance of the two-stream CNN-LSTM architecture is compared with that of single-stream CNN-LSTM models on the same RGB and point cloud datasets, separately. The results demonstrate that the proposed multi-stream approach achieved an accuracy of 94.67%, outperforming existing state-of-the-art single-stream models, which achieved 90.67% accuracy for the RGB-based model and 92.00% for the point cloud-based model. These findings underscore the potential of the proposed activity recognition method, making it highly effective for automatic real-time monitoring of excavator activities, thereby laying the groundwork for future integration into digital twin systems for proactive maintenance and intelligent equipment management. Full article

The response of a temperature sensor in a gas stream depends on several heat transfer phenomena. The temperature of the thermocouple’s hot junction in the hot stream is lower than the measured temperature, which causes a measurement error. Compensation for this error and interpretation of the values indicated by the temperature sensor are possible by using a sensor dynamics model. Changes over time of the hot junction temperature as well as the entire thermocouple temperature in a stream are solved using the finite element method. Fluid flow and heat transfer equations are solved for a particular sensor geometry. This article presents a method for identifying a temperature sensor model using the results of numerical modeling of the response to temperature changes of the fluid stream, in which the input and output signal waveforms are recorded and then used by the estimator of a model coefficient. It is demonstrated that the dynamics of a bare-bead thermocouple sensor are well-described by a first-order transfer function. The proposed method was used to study the influence of stream velocity on the reaction of two sensors differing in the diameter of the wires, and the effect of radiative heat transfer on the model coefficients was examined by enabling and disabling selected models. The results obtained at several calculation points show the influence of the stream outflow velocity and selected geometric parameters on the value of the transfer function coefficients, i.e., transfer function gain and time constant. This study provides quantitative models of changes in sensor dynamics as functions of the coefficients. Full article

The olive oil sector constitutes a fundamental pillar in the Mediterranean region from socio-economic and cultural perspectives. Nonetheless, it produces significant amounts of waste, leading to numerous environmental issues. These waste streams contain valuable compounds that can be recovered and utilized as inputs for various applications. This study introduces a novel value chain for olive wastes, focused on extracting lignin from olive pomace by ionic liquids and polyphenols from olive mill wastewater, which are then incorporated as hybrid nanoparticles in the formulation of an innovative starch-based biofertilizer. This biofertilizer, obtained by using residual wastewater as a source of soluble nitrogen, acting at the same time as a plasticizer for the biopolymer, was demonstrated to surpass traditional NPK biofertilizers’ efficiency, allowing for root growth and foliage in drought conditions. In order to recognize the environmental impact due to its production and align it with the technical output, the circularity and environmental performance of the proposed system were innovatively evaluated through a combination of Life Cycle Assessment (LCA) and the Material Circularity Indicator (MCI). LCA results indicated that the initial upcycling process was potentially characterized by significant hot spots, primarily related to energy consumption (>0.70 kWh/kg of water) during the early processing stages. As a result, the LCA score of this preliminary version of the biofertilizer may be higher than that of conventional commercial products, due to reliance on thermal processes for water removal and the substantial contribution (56%) of lignin/polyphenol precursors to the total LCA score. Replacing energy-intensive thermal treatments with more efficient alternatives represents a critical area for improvement. The MCI value of 0.84 indicates limited potential for further enhancement. Full article

Most industrial processes depend on heat, electricity, demineralized water, and chemical inputs, which themselves are produced through energy- and resource-intensive industrial activities. In this work, acetic acid (AA) production from syngas (CO, CO₂, and H₂) fermentation is explored and compared against a thermochemical fossil benchmark and other thermochemical/biological processes across four main Key Performance Indicators (KPI)—electricity use, heat use, water consumption, and carbon footprint (CF)—for the years 2023 and 2050 in Portugal and France. CF was evaluated through transparent and public inventories for all the processes involved in chemical production and utilities. Spreadsheet-traceable matrices for hotspot identification were also developed. The fossil benchmark, with all the necessary cascade processes, was 0.64 kg CO_2-eq/kg AA, 1.53 kWh/kg AA, 22.02 MJ/kg AA, and 1.62 L water/kg AA for the Portuguese 2023 energy mix, with a reduction of 162% of the CO₂-eq in the 2050 energy transition context. The results demonstrated that industrial practices would benefit greatly from the transition from fossil to renewable energy and from more sustainable chemical sources. For carbon-intensive sectors like steel or cement, the acetogenic syngas fermentation appears as a scalable bridge technology, converting the flue gas waste stream into marketable products and accelerating the transition towards a circular economy. Full article

The integration of traditional geomorphological approaches with advanced artificial intelligence techniques represents a promising frontier in flood risk assessment for arid regions. This study presents a comprehensive analysis of the Wadi Ranuna basin in Medina, Saudi Arabia, combining detailed morphometric parameters with advanced Geospatial Artificial Intelligence (GeoAI) algorithms to enhance flood susceptibility modeling. Using digital elevation models (DEMs) and geographic information systems (GISs), we extracted 23 morphometric parameters across 67 sub-basins and applied XGBoost, Random Forest, and Gradient Boosting (GB) models to predict both continuous flood susceptibility indices and binary flood occurrences. The machine learning models utilize morphometric parameters as input features to capture complex non-linear interactions, including threshold-dependent relationships where the stream frequency impact intensifies above 3.0 streams/km², and the compound effects between the drainage density and relief ratio. The analysis revealed that the basin covers an area of 188.18 km² with a perimeter of 101.71 km and contains 610 streams across six orders. The basin exhibits an elongated shape with a form factor of 0.17 and circularity ratio of 0.23, indicating natural flood-moderating characteristics. GB emerged as the best-performing model, achieving an RMSE of 6.50 and an R² value of 0.9212. Model validation through multi-source approaches, including field verification at 35 locations, achieved 78% spatial correspondence with documented flood events and 94% accuracy for very high susceptibility areas. SHAP analysis identified the stream frequency, overland flow length, and drainage texture as the most influential predictors of flood susceptibility. K-Means clustering uncovered three morphometrically distinct zones, with Cluster 1 exhibiting the highest flood risk potential. Spatial analysis revealed 67% of existing infrastructure was located within high-risk zones, with 23 km of major roads and eight critical facilities positioned in flood-prone areas. The spatial distribution of GBM-predicted flood susceptibility identified high-risk zones predominantly in the central and southern parts of the basin, covering 12.3% (23.1 km²) of the total area. This integrated approach provides quantitative evidence for informed watershed management decisions and demonstrates the effectiveness of combining traditional morphometric analysis with advanced machine learning techniques for enhanced flood risk assessment in arid regions. Full article