Search Results (717)

This study synthesizes the rapidly expanding evidence on how digital technologies reshape international trade, with a particular focus on small and medium-sized enterprises (SMEs). Guided by two research questions—(RQ1) How do digital tools influence the volume and composition of cross-border trade? and (RQ2) How do these effects vary by countries’ development level and firm size?—we conducted a PRISMA-compliant systematic literature review covering 2010–2024. Searches across eight major databases yielded 1857 records; after duplicate removal, title/abstract screening, full-text assessment, and Mixed Methods Appraisal Tool (MMAT 2018) quality checks, 86 peer-reviewed English-language studies were retained. Findings reveal three dominant technology clusters: (1) e-commerce platforms and cloud services, (2) IoT-enabled supply chain solutions, and (3) emerging AI analytics. E-commerce and cloud adoption consistently raise export intensity—doubling it for digitally mature SMEs—while AI applications are the fastest-growing research strand, particularly in East Asia and Northern Europe. However, benefits are uneven: firms in low-infrastructure settings face higher fixed digital costs, and cybersecurity and regulatory fragmentation remain pervasive obstacles. By integrating trade economics with development and SME internationalization studies, this review offers the first holistic framework that links national digital infrastructure and policy support to firm-level export performance. It shows that the trade-enhancing effects of digitalization are contingent on robust broadband penetration, affordable cloud access, and harmonized data-governance regimes. Policymakers should, therefore, prioritize inclusive digital-readiness programs, while business leaders should invest in complementary capabilities—data analytics, cyber-risk management, and cross-border e-logistics—to fully capture digital trade gains. This balanced perspective advances theory and practice on building resilient, equitable digital trade ecosystems. Full article

The acquisition of plant phenotypic traits is essential for selecting superior varieties, improving crop yield, and supporting precision agriculture and agricultural decision-making. Therefore, it plays a significant role in modern agriculture and plant science research. Traditional manual measurements of phenotypic traits are labor-intensive and inefficient. In contrast, combining 3D reconstruction technologies with autonomous vehicles enables more intuitive and efficient trait acquisition. This study proposes a 3D semantic reconstruction system based on an improved ORB-SLAM3 framework, which is mounted on an unmanned vehicle to acquire phenotypic traits in tomato cultivation scenarios. The vehicle is also equipped with the A * algorithm for autonomous navigation. To enhance the semantic representation of the point cloud map, we integrate the BiSeNetV2 network into the ORB-SLAM3 system as a semantic segmentation module. Furthermore, a two-stage filtering strategy is employed to remove outliers and improve the map accuracy, and OctoMap is adopted to store the point cloud data, significantly reducing the memory consumption. A spherical fitting method is applied to estimate the number of tomato fruits. The experimental results demonstrate that BiSeNetV2 achieves a mean intersection over union (mIoU) of 95.37% and a frame rate of 61.98 FPS on the tomato dataset, enabling real-time segmentation. The use of OctoMap reduces the memory consumption by an average of 96.70%. The relative errors when predicting the plant height, canopy width, and volume are 3.86%, 14.34%, and 27.14%, respectively, while the errors concerning the fruit count and fruit volume are 14.36% and 14.25%. Localization experiments on a field dataset show that the proposed system achieves a mean absolute trajectory error (mATE) of 0.16 m and a root mean square error (RMSE) of 0.21 m, indicating high localization accuracy. Therefore, the proposed system can accurately acquire the phenotypic traits of tomato plants, providing data support for precision agriculture and agricultural decision-making. Full article

The quality of wine largely depends on the quality of wine grapes, which is determined by their chemical composition. Therefore, measuring parameters related to grape ripeness, such as soluble solids content (SSC), is crucial for harvesting high-quality grapes. Visible–Near-Infrared (Vis/NIR) spectroscopy enables effective, non-destructive detection of SSC in grapes. However, commercial Vis/NIR spectrometers are often expensive, bulky, and power-consuming, making them unsuitable for on-site applications. This article integrated the AS7265X sensor to develop a low-cost handheld IoT multispectral detection device, which can collect 18 variables in the wavelength range of 410–940 nm. The data can be sent in real time to the cloud configuration, where it can be backed up and visualized. After simultaneously removing outliers detected by both Monte Carlo (MC) and principal component analysis (PCA) methods from the raw spectra, the SSC prediction model was established, resulting in an ${\mathrm{R}}_V^2$ of 0.697. Eight preprocessing methods were compared, among which moving average smoothing (MAS) and Savitzky–Golay smoothing (SGS) improved the ${\mathrm{R}}_V^2$ to 0.756 and 0.766, respectively. Subsequently, feature wavelengths were selected using UVE and SPA, reducing the number of variables from 18 to 5 and 6, respectively, further increasing the ${\mathrm{R}}_V^2$ to 0.809 and 0.795. The results indicate that spectral data optimization methods are effective and essential for improving the performance of SSC prediction models. The IoT Vis/NIR Spectroscopic System proposed in this study offers a miniaturized, low-cost, and practical solution for SSC detection in wine grapes. Full article

This study utilizes UAV-based LiDAR to analyze doline microtopography within a karst mountainous terrain. The study area, ‘Gulneomjae’ in Mungyeong City, South Korea, features steep slopes, limited accessibility, and abundant vegetation—conditions that traditionally hinder accurate topographic surveying. UAV LiDAR data were acquired using the DJI Matrice 300 RTK equipped with a Zenmuse L2 sensor, enabling high-density point cloud generation (98 points/m²). The point clouds were processed to remove non-ground points and generate a 0.25 m resolution DEM using TIN interpolation. A total of seven dolines were detected and delineated, and their morphometric characteristics—including area, perimeter, major and minor axes, and elevation—were analyzed. These results were compared with a 1:5000-scale DEM derived from the 2013 National Basic Map. Visual and numerical comparisons highlighted significant improvements in spatial resolution and feature delineation using UAV LiDAR. Although the 1:5000-scale DEM enables general doline detection, UAV LiDAR facilitates more precise boundary extraction and morphometric analysis. The study demonstrates the effectiveness of UAV LiDAR for detailed topographic mapping in complex karst terrains and offers a foundation for future automated classification and temporal change analysis. Full article

Because neural networks pervade many industrial domains and are increasingly complex and accurate, the trained models themselves have become valuable intellectual properties. Developing highly accurate models demands increasingly higher investments of time, capital, and expertise. Many of these models are commonly deployed in cloud services and on resource-constrained edge devices. Consequently, safeguarding them is critically important. Neural network watermarking offers a practical solution to address this need by embedding a unique signature, either as a hidden bit-string or as a distinctive response to specially crafted “trigger” inputs. This allows owners to subsequently prove model ownership even if an adversary attempts to remove the watermark through attacks. In this manuscript, we adapt three state-of-the-art watermarking methods to “tiny” neural networks deployed on edge platforms by exploiting symmetry-related properties that ensure robustness and efficiency. In the context of machine learning, “tiny” is broadly used as a term referring to artificial intelligence techniques deployed in low-energy systems in the mW range and below, e.g., sensors and microcontrollers. We evaluate the robustness of the selected techniques by simulating attacks aimed at erasing the watermark while preserving the model’s original performances. The results before and after attacks demonstrate the effectiveness of these watermarking schemes in protecting neural network intellectual property without degrading the original accuracy. Full article

The current methods for removing thick clouds from remote-sensing images face significant limitations, including the integration of thick cloud images with synthetic aperture radar (SAR) ground information, the provision of meaningful guidance for SAR ground data, and the accurate reconstruction of textures in cloud-covered regions. To overcome these challenges, we introduce SAR-DeCR, a novel method for thick cloud removal in satellite remote-sensing images. SAR-DeCR utilizes a diffusion model combined with the transformer architecture to synthesize accurate texture details guided by SAR ground information. The method is structured into three distinct phases: coarse cloud removal (CCR), SAR-Fusion (SAR-F) and cloud-free diffusion (CF-D), aimed at enhancing the effectiveness of the thick cloud removal. In CCR, we significantly employ the transformer’s capability for long-range information interaction, which significantly strengthens the cloud removal process. In order to overcome the problem of missing ground information after cloud removal and ensure that the ground information produced is consistent with SAR data, we introduced SAR-F, a module designed to incorporate the rich ground information in synthetic aperture radar (SAR) into the output of CCR. Additionally, to achieve superior texture reconstruction, we introduce prior supervision based on the output of the coarse cloud removal, using a pre-trained visual-text diffusion model named cloud-free diffusion (CF-D). This diffusion model is encouraged to follow the visual prompts, thus producing a visually appealing, high-quality result. The effectiveness and superiority of SAR-DeCR are demonstrated through qualitative and quantitative experiments, comparing it with other state-of-the-art (SOTA) thick cloud removal methods on the large-scale SEN12MS-CR dataset. Full article

Optical remote sensing images suffer from information loss due to cloud interference, while Synthetic Aperture Radar (SAR), capable of all-weather and day–night imaging capabilities, provides crucial auxiliary data for cloud removal and reconstruction. However, existing cloud removal methods face the following key challenges: insufficient utilization of spatiotemporal information in multi-temporal data, and fusion challenges arising from fundamentally different imaging mechanisms between optical and SAR images. To address these challenges, a spatiotemporal feature interaction-based cloud removal method is proposed to effectively fuse SAR and optical images. Built upon a conditional generative adversarial network framework, the method incorporates three key modules: a multi-temporal spatiotemporal feature joint extraction module, a spatiotemporal information interaction module, and a spatiotemporal discriminator module. These components jointly establish a many-to-many spatiotemporal interactive learning network, which separately extracts and fuses spatiotemporal features from multi-temporal SAR–optical image pairs to generate temporally consistent, cloud-free image sequences. Experiments on both simulated and real datasets demonstrate the superior performance of the proposed method. Full article

In target detection tasks equipped with depth sensors, it is crucial to adopt the point cloud pretreatment process, which is directly related to the quality of the obtained three-dimensional model of the target. However, there are few methods that can be combined with common preprocessing methods to quickly process ToF camera output. In real-life experiments, the common method is to adopt multiple types of preprocessing methods and adjust parameters separately. We proposed PcBD, a method that integrates outlier removal, boundary detection, and smooth sliders. PcBD does not limit the number of input points, and can remove outliers and predict smooth projection boundaries at one time while ensuring that the total number of points remains unchanged. We also introduced Bound57, a benchmark dataset that contains point clouds with synthetic noise, outliers, and projected boundary labels. Experimental results show that PcBD performs significantly better than state-of-the-art methods in various de-noising and boundary detection tasks. Full article

To address the challenges of dynamic object interference and redundant information representation in map construction for indoor dynamic environments, this paper proposes an indoor dynamic environment mapping method based on semantic fusion and hierarchical filtering. First, prior dynamic object masks are obtained using the YOLOv8 model, and geometric constraints between prior static objects and dynamic regions are introduced to identify non-prior dynamic objects, thereby eliminating all dynamic features (both prior and non-prior). Second, an initial semantic point cloud map is constructed by integrating prior static features from a semantic segmentation network with pose estimates from an RGB-D camera. Dynamic noise is then removed using statistical outlier removal (SOR) filtering, while voxel filtering optimizes point cloud density, generating a compact yet texture-rich semantic dense point cloud map with minimal dynamic artifacts. Subsequently, a multi-resolution semantic octree map is built using a recursive spatial partitioning algorithm. Finally, point cloud poses are corrected via Transform Frame (TF) transformation, and a 2D traversability grid map is generated using passthrough filtering and grid projection. Experimental results demonstrate that the proposed method constructs multi-level semantic maps with rich information, clear structure, and high reliability in indoor dynamic scenarios. Additionally, the map file size is compressed by 50–80%, significantly enhancing the reliability of mobile robot navigation and the efficiency of path planning. Full article

To address the inefficiencies and limited spatial resolution of traditional single-point monitoring techniques, this study proposes a multi-scale analysis method that integrates the Multi-Scale Model-to-Model Cloud Comparison (M3C2) algorithm with least-squares plane fitting. This approach employs the M3C2 algorithm for qualitative full-field deformation detection and utilizes least-squares plane fitting for quantitative feature extraction. When applied to the approach span of a cross-river bridge in Hubei Province, China, this method leverages dense point clouds (greater than 500 points per square meter) acquired using a Leica RTC360 scanner. Data preprocessing incorporates curvature-adaptive cascade denoising, achieving over 98% noise removal while retaining more than 95% of structural features, along with octree-based simplification. By extracting multi-level slice features from bridge decks and piers, this method enables the simultaneous analysis of global trends and local deformations. The results revealed significant deformation, with an average settlement of 8.2 mm in the left deck area. The bridge deck exhibited a deformation trend characterized by left and higher right in the vertical direction, while the bridge piers displayed noticeable tilting, particularly with the maximum offset of the rear pier columns reaching 182.2 mm, which exceeded the deformation of the front pier. The bridge deck’s micro-settlement error was ±1.2 mm, and the pier inclination error was ±2.8 mm, meeting the Chinese Highway Bridge Maintenance Code (JTG H11-2004) and the American Association of State Highway and Transportation Officials (AASHTO) standards, and the multi-scale algorithm achieved engineering-level accuracy. Utilizing point cloud densities >500 pt/m², the M3C2 algorithm achieved a spatial resolution of 0.5 mm, enabling sub-millimeter full-field analysis for complex scenarios. This method significantly enhances bridge safety monitoring precision, enhances the precision of intelligent systems monitoring, and supports the development of targeted systems as pile foundation reinforcement efforts and as improvements to foundations. Full article

Reliable soil moisture projections are critical for optimizing crop productivity and water savings in irrigation in arid and semi-arid regions. However, capturing their spatial and temporal variability is difficult when using individual observations, modeling, or satellite-based methods. Here, we present an integrated framework that combines satellite-derived soil moisture estimates, ground-based observations, the HYDRUS-1D vadose zone model, and the ensemble Kalman filter (EnKF) data assimilation method to improve soil moisture simulations over saline-affected farmland in the Hetao irrigation district. Vegetation effects were first removed using the water cloud model; after correction, a cubic regression using the vertical transmit/vertical receive (VV) signal retrieved surface moisture with an R² value of 0.7964 and a root mean square error (RMSE) of 0.021 cm³·cm⁻³. HYDRUS-1D, calibrated against multi-depth field data (0–80 cm), reproduced soil moisture profiles at 17 sites with RMSEs of 0.017–0.056 cm³·cm⁻³. The EnKF assimilation of satellite and ground observations further reduced the errors to 0.008–0.017 cm³·cm⁻³, with the greatest improvement in the 0–20 cm layer; the accuracy declined slightly with depth but remained superior to either data source alone. Our study improves soil moisture simulation accuracy and closes the knowledge gaps in multi-source data integration. This framework supports sustainable land management and irrigation policy in vulnerable farming regions. Full article

This study investigates the impact of desert dust on precipitation patterns using multi-model simulations. Dust-based processes of formation/removal of ice nuclei (IN) and cloud condensation nuclei (CCN) are investigated by using both the online access model WRF-CHIMERE and the online integrated model WRF-Chem. Comparisons of model predictions with rainfall measurements (GRISO: Spatial Interpolation Generator from Rainfall Observations) over the Italian peninsula show the models’ ability to reproduce heavy orographic precipitation in alpine regions. To quantify the impact of the mineral dust transport concomitant to the atmospheric river (AR) on cloud formation, a sensitivity study is performed by using the WRF-CHIMERE model (i) by setting dust concentrations to zero and (ii) by modifying the settings of the Thompson Aerosol-Aware microphysics scheme. Statistical comparisons revealed that WRF-CHIMERE outperformed WRF-Chem. It achieved a correlation coefficient of up to 0.77, mean bias (MB) between +3.56 and +5.01 mm/day, and lower RMSE and MAE values (~32 mm and ~22 mm, respectively). Conversely, WRF-Chem displayed a substantial underestimation, with an MB of −25.22 mm/day and higher RMSE and MAE values. Our findings show that, despite general agreement in spatial precipitation patterns, both models significantly underestimated the peak daily rainfall in pre-alpine regions (e.g., 216 mm observed at Malga Valine vs. 130–140 mm simulated, corresponding to a 35–40% underestimation). Although important instantaneous changes in precipitation and temperature were modeled at a local scale, no significant total changes in precipitation or air temperature averaged over the entire domain were observed. These results underline the complexity of aerosol–cloud interactions and the need for improved parameterizations in coupled meteorological models. Full article

The removal of per- and polyfluoroalkyl substances (PFAS) from global aquatic environments is an emerging issue. However, little attention has been paid to addressing accumulated PFAS through their removal. This study demonstrates the encapsulation of perfluoroalkyl carboxylic acids (PFCAs) within polymer microspheres that dissolve in supercritical carbon dioxide (scCO₂). PFCAs were effectively captured by a hindered amine-supported monomer, 2,2,6,6-tetramethyl-4-piperidyl methacrylate (TPMA), in methanol (MeOH) through a simple acid-base reaction. The PFCA-loaded TPMA underwent dispersion polymerization in MeOH in the presence of poly(N-vinylpyrrolidone) (PVP) as a surfactant, producing microspheres with high monomer conversions. The microsphere size depended on the molecular weight and concentration of PVP, as well as the perfluoroalkyl chain length of the PFCAs. X-ray photoelectron spectroscopy (XPS) revealed that the perfluoroalkyl chains migrated from the interior to the surface of the microspheres when exposed to air. These surface perfluoroalkyl chains facilitated dissolution of the microspheres in scCO₂, with cloud points observed under relatively mild conditions. These findings suggest the potential for managing PFCA-encapsulated microspheres in the scCO₂ phase deep underground via CO₂ sequestration. Full article

The use of UAVs for analyzing soil degradation processes, particularly erosion, has become a crucial tool in environmental monitoring. However, the use of LiDAR (Light Detection and Ranging) or TLS (Terrestrial Lasser Scanner) may not be affordable for many researchers because of the elevated costs and difficulties for cloud processing to present a valuable option for rapid landscape assessment following extreme events like Mediterranean storms. This study focuses on the application of drone-based remote sensing with only an RGB camera in geomorphological mapping. A key objective is the removal of vegetation from imagery to enhance the analysis of erosion and sediment transport dynamics. The research was carried out over a cereal cultivation plot in Málaga Province, an area recently affected by high-intensity rainfalls exceeding 100 mm in a single day in the past year, which triggered significant soil displacement. By processing UAV-derived data, a Digital Elevation Model (DEM) was generated through geostatistical techniques, refining the Digital Surface Model (DSM) to improve topographical change detection. The ability to accurately remove vegetation from aerial imagery allows for a more precise assessment of erosion patterns and sediment redistribution in geomorphological features with rapid spatiotemporal changes. Full article

To assess the condition of cultural heritage assets for conservation, reality-based 3D models can be analyzed using FEA (finite element analysis) software, yielding valuable insights into their structural integrity. Three-dimensional point clouds obtained through photogrammetric and laser scanning techniques can be transformed into volumetric data suitable for FEA by utilizing voxels. When directly using the point cloud data in this process, it is crucial to employ the highest level of accuracy. The fidelity of r point clouds can be compromised by various factors, including uncooperative materials or surfaces, poor lighting conditions, reflections, intricate geometries, and limitations in the precision of the instruments. This data not only skews the inherent structure of the point cloud but also introduces extraneous information. Hence, the geometric accuracy of the resulting model may be diminished, ultimately impacting the reliability of any analyses conducted upon it. The removal of noise from point clouds, a crucial aspect of 3D data processing, known as point cloud denoising, is gaining significant attention due to its ability to reveal the true underlying point cloud structure. This paper focuses on evaluating the geometric precision of the voxelization process, which transforms denoised 3D point clouds into volumetric models suitable for structural analyses. Full article