Search Results (13)

In the context of Industry 4.0, reliable automatic inspection of weld surface defects is critical for structural safety, yet current deep learning-based detectors struggle with the extreme scale variation and anisotropic shapes characteristic of weld flaws such as pores, cracks, and lack of fusion. Existing YOLO-family models, although effective on general-purpose datasets, often fail to robustly localize tiny defects and long, slender discontinuities while remaining lightweight enough for industrial edge deployment. A critical research gap lies in the lack of task-specific optimization for weld defects: standard attention mechanisms are isotropic and cannot capture linear defect continuity, while existing loss functions ignore scale disparity between tiny pores (area < 100 pixels²) and large incomplete fusion defects (area > 5000 pixels²), leading to unstable regression. Here, we propose a dual-optimized lightweight YOLOv11 framework tailored for weld defect detection that addresses both feature representation and bounding-box regression. Here, we propose a dual-optimized lightweight YOLOv11 framework tailored for weld defect detection that addresses both feature representation and bounding-box regression. First, we introduce WeldSimAM, an enhanced attention module that augments parameter-free SimAM with directional (horizontal/vertical) and channel-wise enhancement to better capture the directional texture of linear weld defects. Second, we develop an Enhanced Normalized Wasserstein Distance (EnNWD) loss, which incorporates scale-disparity penalties and relative-area-based weighting to mitigate sample imbalance and improve regression accuracy for tiny and large-aspect-ratio targets. Validated via 10-fold cross-validation on three datasets (self-built + two public), the method achieves 99.48% mAP@0.5 and 73.29% mAP@0.5:0.95, outperforming YOLOv11 by 0.13 and 3.76 percentage points (p < 0.01, two-tailed t-test), with 5.21 MB and 132 FPS on NVIDIA RTX 4090. It also surpasses non-YOLO SOTA methods (e.g., EfficientDet-Lite3) by 3.8–5.5 percentage points in mAP@0.5 (p < 0.05), offering a practical real-time solution for industrial inspection. Full article

Shale is a common rock in oil and gas extraction, and the study of its nonlinear mechanical behavior is crucial for the development of engineering techniques such as hydraulic fracturing. This paper establishes a new coupled elastic–plastic damage model based on the second law of thermodynamics, the strain equivalence principle, the non-associated flow rule, and the Drucker–Prager yield criterion. This model is used to describe the mechanical behavior of shale before and after peak strength and has been implemented in ABAQUS via UMAT for numerical computation. The model comprehensively considers the quasi-brittle and anisotropic characteristics of shale, as well as the strength degradation caused by damage during both the elastic and plastic phases. A damage yield function has been established as a criterion for damage occurrence, and the constitutive integration algorithm has been derived using a regression mapping algorithm. Compared with experimental data from La Biche shale in Canada, the theoretical model accurately simulated the stress–strain curves and volumetric–axial strain curves of shale under confining pressures of 5 MPa, 25 MPa, and 50 MPa. When compared with experimental data from shale in Western Hubei and Eastern Chongqing, China, the model precisely fitted the stress–strain curves of shale at pressures of 30 MPa, 50 MPa, and 70 MPa, and at bedding angles of 0°, 22.5°, 45°, and 90°. This proves that the model can effectively predict the failure behavior of shale under different confining pressures and bedding angles. Additionally, a sensitivity analysis has been performed on parameters such as the plastic hardening rate b, damage evolution rate B_ω, weighting factor r, and damage softening parameter a. This research is expected to provide theoretical support for the efficient extraction technologies of shale oil and gas. Full article

In recent years, the advancement of CubeSat technology has led to the emergence of high-resolution, flexible imaging satellites as a pivotal source of information for the efficient and precise monitoring of crops. However, the dynamic geometry inherent in flexible side-view imaging poses challenges in acquiring the high-precision reflectance data necessary to accurately retrieve crop parameters. This study aimed to develop an angular correction method designed to generate nadir reflectance from high-resolution satellite side-swing imaging data. The method utilized the Anisotropic Flat Index (AFX) in conjunction with a fixed set of Bidirectional Reflectance Distribution Function (BRDF) parameters to compute the nadir reflectance for the Jilin-1 GP01/02 multispectral imager (PMS). Crop parameter retrieval was executed using regression models based on vegetation indices, the leaf area index (LAI), fractional vegetation cover (FVC), and chlorophyll (T850 nm/T720 nm) values estimated based on angle corrected reflectance compared with field measurements taken in the Inner Mongolia Autonomous Region. The findings demonstrate that the proposed angular correction method significantly enhances the retrieval accuracy of the LAI, FVC, and chlorophyll from Jilin-1 GP01/02 PMS data. Notably, the retrieval accuracy for the LAI and FVC improved by over 25%. We expect that this approach will exhibit considerable potential to improve crop monitoring accuracy from high-resolution satellite side-view imaging data. Full article

For nonparametric regression estimation, conventional research all focus on isotropic regression function. In this paper, a linear wavelet estimator of anisotropic regression function is constructed, the rate of convergence of this estimator is discussed in anisotropic Besov spaces. More importantly, in order to obtain an adaptive estimator, a regression estimator is proposed with scaling parameter data-driven selection rule. It turns out that our results attain the optimal convergence rate of nonparametric pointwise estimation. Full article

Experimental flow stress–strain data under different stress states are often used to calibrate the plastic constitutive model of anisotropic metal materials or identify the appropriate model that is able to reproduce their plastic deformation behavior. Since the experimental stress–strain data are discrete, they need to be mathematically returned to a continuous function to be used to describe an equivalent hardening increment. However, the regression results obtained using existing regression models are not always accurate, especially for stress–strain curves under biaxial stress loading conditions. Therefore, a new regression model is proposed in this paper. The highest-order term in the recommended form of the new model is quadratic, so the functional relationships between stress–strain components can be organized into explicit expressions. All the experimental data of the uniform deformation stage can be substituted into the new model to reasonably reproduce the biaxial experimental stress–strain data. The regression results of experimental data show that the regression accuracy of the new model is greatly improved, and the residual square sum SSE of the regression curves of the new model reduced to less than 50% of the existing three models. The regression results of stress–strain curves show significant differences in describing the yield and plastic flow characteristics of anisotropic metal materials, indicating that accurate regression results are crucial for accurately describing the anisotropic yielding and plastic flow behaviors of anisotropic metal materials. Full article

The aim of the present work is to investigate the efficiency of flat-plate solar panels in Greece for delivering solar energy. In this study, the solar panels are mounted on a two-axis tracker, which follows the daily path of the sun. In this context, the annual energy sums are estimated on such surfaces from hourly solar horizontal radiation values at forty-three locations, covering all of Greece. The solar horizontal radiation values are embedded in the typical meteorological years of the sites obtained from the PVGIS tool. All calculations use near-real surface-albedo values for the sites, and isotropic and anisotropic models are used to estimate the diffuse-inclined radiation. The analysis provides non-linear regression expressions for the energy sums as a function of time (month, season). The annual energy sums are found to vary between 2247 kWhm⁻² and 2878 kWhm⁻² under all-sky conditions with the anisotropic transposition model. Finally, maps of Greece showing the distribution of the annual and seasonal solar energy sums under all- and clear-sky conditions are derived for the first time, and these maps constitute the main innovation of this work. Full article

Background: Education is believed to contribute positively to brain structure and function, as well as to cognitive reserve. One of the brain regions most impacted by education is the medial temporal lobe (MTL), a region that houses the hippocampus, which has an important role in learning processes and in consolidation of memories, and is also known to undergo neurogenesis in adulthood. We aimed to investigate the influence of education on the absolute cell numbers of the MTL (comprised by the hippocampal formation, amygdala, and parahippocampal gyrus) of men without cognitive impairment. Methods: The Isotropic Fractionator technique was used to allow the anisotropic brain tissue to be transformed into an isotropic suspension of nuclei, and therefore assess the absolute cell composition of the MTL. We dissected twenty-six brains from men aged 47 to 64 years, with either low or high education. Results: A significant difference between groups was observed in brain mass, but not in MTL mass. No significant difference was found between groups in the number of total cells, number of neurons, and number of non-neuronal cells. Regression analysis showed that the total number of cells, number of neurons, and number of non-neuronal cells in MTL were not affected by education. Conclusions: The results indicate a resilience of the absolute cellular composition of the MTL of typical men to low schooling, suggesting that the cellularity of brain regions is not affected by formal education. Full article

Mechanical properties of fiber-reinforced polymers are sensitive to environmental influences due to the presence of the polymer matrix but inhomogeneous and anisotropic due to the presence of the fibers. Hence, structural analysis with mechanical properties as a function of loading, environment, design, and material condition produces more precise, reliable, and economic structures. In the present study, an analytical model is developed that can predict engineering values as well as non-linear stress–strain curves as a function of six independent parameters for short fiber-reinforced polymers manufactured by injection molding. These parameters are the strain, temperature, humidity, fiber content, fiber orientation, and thickness of the specimen. A three-point test matrix for each independent parameter is used to obtain experimental data. To insert the effect of in-homogenous and anisotropic distribution of fibers in the analytical model, microCT analysis is done. Similarly, dynamic mechanical thermal analysis (DMTA) is done to insert the viscoelastic effect of the material. The least mean square regression method is used to predict empirical formulas. The standard error of regression for the fitting of the model with experimental stress–strain curves is closely controlled below 2% of the stress range. This study provides user-specific material data for simulations with specific material, loading, and environmental conditions. Full article

This study presents a new method that mitigates biases between the normalized difference vegetation index (NDVI) from geostationary (GEO) and low Earth orbit (LEO) satellites for Earth observation. The method geometrically and spectrally transforms GEO NDVI into LEO-compatible GEO NDVI, in which GEO’s off-nadir view is adjusted to a near-nadir view. First, a GEO-to-LEO NDVI transformation equation is derived using a linear mixture model of anisotropic vegetation and nonvegetation endmember spectra. The coefficients of the derived equation are a function of the endmember spectra of two sensors. The resultant equation is used to develop an NDVI transformation method in which endmember spectra are automatically computed from each sensor’s data independently and are combined to compute the coefficients. Importantly, this method does not require regression analysis using two-sensor NDVI data. The method is demonstrated using Himawari 8 Advanced Himawari Imager (AHI) data at off-nadir view and Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) data at near-nadir view in middle latitude. The results show that the magnitudes of the averaged NDVI biases between AHI and MODIS for five test sites (0.016–0.026) were reduced after the transformation (<0.01). These findings indicate that the proposed method facilitates the combination of GEO and LEO NDVIs to provide NDVIs with smaller differences, except for cases in which the fraction of vegetation cover (FVC) depends on the view angle. Further investigations should be conducted to reduce the remaining errors in the transformation and to explore the feasibility of using the proposed method to predict near-real-time and near-nadir LEO vegetation index time series using GEO data. Full article

Various crystallite size estimation methods were used to analyze X-ray diffractograms of spherical cerium dioxide and titanium dioxide anatase nanoparticles aiming to evaluate their reliability and limitations. The microstructural parameters were estimated from several integral breadth methods such as Scherrer, Monshi, Williamson–Hall, and their variants: (i) uniform deformation model, (ii) uniform strain deformation model, and (iii) uniform deformation energy density model. We also employed the size–strain plot and Halder–Wagner method. For this purpose, an instrumental resolution function of an Al₂O₃ standard was used to subtract the instrumental broadening to estimate the crystallite sizes and strain, and the linear regression analysis was used to compare all the models based on the coefficient of determination. The Rietveld whole powder pattern decomposition method was introduced for comparison purposes, being the best candidate to fit the X-ray diffraction data of metal-oxide nanoparticles. Refined microstructural parameters were obtained using the anisotropic spherical harmonic size approach and correlated with the above estimation methods and transmission electron microscopy images. In addition, μ-Raman spectra were recorded for each material, estimating the mean crystallite size for comparison by means of a phonon confinement model. Full article

The ¹H chemical shielding anisotropy (CSA) is an NMR parameter that is exquisitely sensitive to the local environment of protons in crystalline systems, but it is difficult to obtain it experimentally due to the need to concomitantly suppress other anisotropic interactions in the solid-state NMR (SSNMR) pulse sequences. The SSNMR measurements of the ¹H CSA are particularly challenging if the fast magic-angle-spinning (MAS) is applied. It is thus important to confront the results of both the single-crystal (SC) and fast-MAS experiments with their theoretical counterparts. Here the plane-waves (PW) DFT calculations have been carried out using two functionals in order to precisely characterize the structures and the ¹H NMR chemical shielding tensors (CSTs) of the solid forms of maleic, malonic, and citric acids, and of L-histidine hydrochloride monohydrate. The level of agreement between the PW DFT and either SC or fast-MAS SSNMR ¹H CSA data has been critically compared. It has been found that for the eigenvalues of the ¹H CSTs provided by the fast-MAS measurements, an accuracy limit of current PW DFT predictions is about two ppm in terms of the standard deviation of the linear regression model, and sources of this error have been thoroughly discussed. Full article

The resolution of the products of numerical weather prediction is limited by the resolution of numerical models and computing resources, which can be improved accurately by a well-chosen interpolation algorithm. This paper is intended to improve the accuracy of spatial interpolation towards wind fields. A new composited multi-scale anisotropic kernel function for weather processes using two-dimensional space information is proposed. To fix the underfitting in this kernel caused by unilateral space information, multiple variables (wind direction, air temperature, and atmospheric pressure) are introduced, which generates a multivariate correction model based on the novel kernel function and Gaussian process regression. Focusing on different weather processes, two multivariate correction models are designed. The new models pave a new way to employ multi-scale local information, and extract the anisotropy and structure information. The experiments on 10 m wind fields for the weather processes without cyclones and for the weather processes with cyclones validate the efficiency. The mean RMSE of the multivariate correction model for the weather processes without cyclones is reduced by around 15% for the u wind component compared with that of a simple composited kernel. For the weather processes with cyclones, the mean RMSE of the novel model declines by around 55% compared to that of spline, and by about 95% compared to that of back propagation neural networks. Full article

Evaluation of the mechanical properties of arterial wall components is necessary for establishing a precise mechanical model applicable in various physiological and pathological conditions, such as remodeling. In this contribution, a new approach for the evaluation of the mechanical properties of aortic media accounting for the lamellar structure is proposed. We assumed aortic media to be composed of two sets of concentric layers, namely sheets of elastin (Layer I) and interstitial layers composed of mostly collagen bundles, fine elastic fibers and smooth muscle cells (Layer II). Biaxial mechanical tests were carried out on human thoracic aortic samples, and histological staining was performed to distinguish wall lamellae for determining the dimensions of the layers. A neo-Hookean strain energy function (SEF) for Layer I and a four-parameter exponential SEF for Layer II were allocated. Nonlinear regression was used to find the material parameters of the proposed microstructural model based on experimental data. The non-linear behavior of media layers confirmed the higher contribution of elastic tissue in lower strains and the gradual engagement of collagen fibers. The resulting model determines the nonlinear anisotropic behavior of aortic media through the lamellar microstructure and can be assistive in the study of wall remodeling due to alterations in lamellar structure during pathological conditions and aging. Full article