Search Results (47)

Target detection in synthetic aperture radar (SAR) images is constrained by three primary challenges. First, speckle noise overlaps heavily with the high-frequency features of target edges in the frequency domain, so standard convolutions cannot suppress noise without sacrificing edge texture. Second, the scattering signature of a SAR target varies markedly with viewing angle, and a fixed-parameter convolution kernel cannot accommodate this spatial non-stationarity. Third, deep and shallow levels of the feature pyramid differ in semantics and resolution, and a naive element-wise sum either introduces noise interference or loses small-target signals. We propose the Frequency–Spatial Detection Network (FSDNet), whose core FSDBlock cascades three operators to address these failure modes in turn. Wavelet Convolution (WTConv) projects features into Haar sub-bands and applies independent low- and high-frequency kernels prior to inverse-DWT reconstruction, suppressing noise while preserving edges. Receptive-Field Attention Convolution (RFAConv) generates location-conditional kernels and so adapts to non-stationary scattering. Spatial Context Self-Attention (SCSA) aggregates discrete scattering points into coherent target representations via long-range grouped attention. At the fusion stage, CGAFusion replaces FPN element-wise addition with a channel–spatial–pixel triple-attention soft switch that mitigates deep–shallow semantic mismatch. On HRSID, FSDNet attains mAP₅₀ $= 92.3$% and mAP_50:95 $= 68.6$%. On SSDD, it attains mAP₅₀ $= 98.7$% and mAP_50:95 $= 74.2$%. Both sets of results consistently surpass the baseline methods. Against the strongest YOLO baseline (YOLOv11n), FSDNet improves HRSID mAP₅₀ by $+1.7$ percentage points (pp) and mAP_50:95 by $+2.3$ pp, and SSDD mAP₅₀ by $+0.5$ pp and mAP_50:95 by $+2.7$ pp; against the capacity-fair YOLOv11s reference (∼51% more parameters), FSDNet still leads on mAP₅₀, mAP_50:95, recall, and F1. Ablation studies and power-spectral-density analyses corroborate the contribution of each module and confirm WTConv’s role in preserving high-frequency target features. Full article

The rapid advancement of payment technologies and potential disintermediation pressure make it important to monitor how actively commercial banks participate in payment and settlement systems. This study conceptualizes bank participation as a multidimensional construct and develops a Bank Payment Participation Index (BPPI or ANR BPPI) using publicly available Reserve Bank of India data for 2011–2012 to 2022–2023. BPPI integrates Financial Capacity (FC), Technological Readiness (TR), Payment Performance (PI), and a PPI-based Technological Advancement/Disintermediation proxy (TAD). TAD, measured as the share of PPI transactions in total payment volumes, enters the index as (1−TAD) because rising non-bank payment penetration reduces banks’ intermediation share; a higher TAD represents a structural drag on bank payment participation, and (1−TAD) converts this drag into a participation-compatible scale. The index applies min–max normalisation, equal-weighted sub-index aggregation, and geometric mean composition with lagged input dimensions. Computations show that the four-component BPPI rises from 230.794 in 2013–2014 to 797.453 in 2022–2023, indicating a strong long-run increase in banking-system participation. The BPPI remains strongly associated with GDP over the 2013–2014 to 2022–2023 sample, with R Square = 0.906 and adjusted R Square = 0.894. Diagnostic tests indicate that the validation is best interpreted as association-based evidence rather than causal proof. The BPPI is proposed as a decomposable monitoring and diagnostic framework that equips regulators and banks to track participation trends and detect structural vulnerabilities over time, subject to future refinement using fixed policy goalposts, bank-level data, and CBDC-specific transaction data. In its present form, the BPPI constitutes a model-stage prototype framework subject to future operationalisation with fixed expert-determined benchmarks and bank-level disaggregated data. Full article

Blade Tip Timing (BTT) is an essential non-contact technique for monitoring vibrations in rotating machinery, but its practical accuracy is often degraded by noise, undersampling, and spectral leakage. This paper proposes a multi-stage robust Bayesian high-resolution identification framework that systematically addresses these challenges. A recursive digital algorithm based on Kalman filtering estimates the rotational speed without requiring once-per-revolution probes, effectively suppressing sensor noise. An attention-enhanced dynamic convolutional autoencoder then generates channel-specific window functions to minimize spectral leakage. The core identification algorithm extracts phases via all-phase FFT and employs sub-bin interpolation to overcome the resolution limitation of conventional FFT. A Tukey-biweight-based robust aggregation strategy is used to suppress the influence of abnormal or unequal-quality sensor channels during multi-channel phase fusion. A Bayesian prior distribution over the vibration order guides the estimation toward physically plausible values under noisy conditions. Finally, a coarse-to-fine multi-stage search strategy drastically reduces computational burden while preserving accuracy. Experiments on a rotor-blade test bench at constant and variable speeds show that the method reduces the noise floor by about 60 dB, achieves a maximum frequency identification error of 7.84%, and accelerates the search by approximately 48.6% compared to exhaustive search. The proposed method provides a reliable and efficient solution for blade health monitoring. Full article

The crude oil market has obvious financial entropy, and there are characteristics such as continuous uncertainty, multi-scale fluctuations and nonlinear state transitions. These characteristics bring challenges to the traditional prediction method. In this context, in order to improve the accuracy of energy financial market prediction, this study proposes an artificial intelligence-driven hybrid prediction framework, ALA-VMD-CASA. This framework is divided into three stages. First, with the goal of minimizing envelope entropy, ALA is introduced to adaptively optimize the hyperparameters of VMD, so as to generate informative sub-modes with reduced entropy. Next, the parallel prediction of each sub-mode is carried out by using the score attention mechanism based on the CNN autoencoder, and its linear time complexity can capture volatility clustering and sudden price fluctuations. Finally, the final price prediction is generated through the aggregation component. The empirical experiment of Brent crude oil spot prices from 2010 to 2025 shows that the ALA-VMD-CASA framework is superior to benchmark models such as ARIMA, RW, RWWD, LSTM, GRU, Transformer and Informer. Compared with the best standalone model, the proposed framework reduces the mean square error by more than 63% and obtains a perfect win rate in expanding-window evaluations. These results prove that the proposed framework is effective and robust for modeling financial entropy and improving energy price forecasting. Full article

Optical and synthetic aperture radar (SAR) image registration faces challenges from nonlinear radiometric distortions and geometric deformations caused by different imaging mechanisms. This paper proposes a coarse-to-fine registration algorithm integrating geographic information constraints with cross-modal feature consistency mapping. The coarse stage employs imaging geometry-based coordinate transformation with airborne navigation data to eliminate scale and rotation differences. The fine stage constructs a multi-scale phase congruency-based feature response aggregation model combined with rotation-invariant descriptors and global-to-local search for sub-pixel alignment. Experiments on integrated airborne optical/SAR datasets demonstrate superior performance with an average RMSE of 2.00 pixels, outperforming both traditional handcrafted methods (3MRS, OS-SIFT, POS-GIFT, GLS-MIFT) and state-of-the-art deep learning approaches (SuperGlue, LoFTR, ReDFeat, SAROptNet) while reducing execution time by 37.0% compared with the best-performing baseline. The proposed coarse registration also serves as an effective preprocessing module that improves SuperGlue’s matching rate by 167% and LoFTR’s by 109%, with a hybrid refinement strategy achieving 1.95 pixels RMSE. The method demonstrates robust performance under challenging conditions, enabling real-time UAV-based multi-sensor fusion applications. Full article

Accurate ultra-short-term wind power forecasting (WPF) is essential for maintaining power grid stability and minimizing economic risks, yet the inherent volatility of wind speed poses significant modeling challenges. To address this, this study proposes an ensemble framework integrating an Improved Triangular Topology Aggregation Optimizer (ITTAO) and a high-frequency adaptive weighting strategy. Methodologically, the ITTAO incorporates multi-strategy mechanisms to overcome the premature convergence of the traditional TTAO, thereby enabling precise hyperparameter optimization for the variational mode decomposition (VMD) and BiLSTM networks. Furthermore, in the reconstruction stage, a dynamic weighting strategy is introduced to modulate the contribution of high-frequency sub-sequences, thereby enhancing the capture of rapid fluctuations. Experimental results across multi-seasonal datasets demonstrate that the proposed hybrid model consistently outperforms representative baselines. Notably, in the most volatile scenarios, the model achieves an NMAE of 1.33%, an NRMSE of 2.20%, and an $R^2$ of 98.18%. The results demonstrate that the proposed model achieves superior forecasting accuracy, enhancing the operational stability of wind farms and the secure integration of wind energy into the power grid. Full article

Monitoring bottom-hole pressure (BHP) is critical for reservoir management and flow assurance, especially in offshore fields where challenging conditions and production losses are more impactful. However, reliability issues and high installation costs of Permanent Downhole Gauges (PDGs) often limit access to this vital data. Soft sensors offer a cost-effective and reliable alternative, serving as backups or replacements for physical sensors. This study proposes a novel data-driven methodology for estimating flowing BHP using wellhead and topside measurements from plant monitoring systems. The framework employs ensemble methods combined with clustering techniques to partition datasets, enabling tailored supervised training for diverse production conditions. Aggregating results from sub-models enhances performance, even with simpler machine learning algorithms. We evaluated Linear Regression, Neural Networks, and Gradient Boosting (XGBoost and LightGBM) as base models. A case study of a Brazilian Pre-Salt offshore oilfield, using data from 60 wells across nine platforms, demonstrated the methodology’s effectiveness. Error metrics remained consistently below 2% across varying production conditions and reservoir lifecycle stages, confirming its reliability. This solution provides a practical, economical alternative for studies and monitoring in wells lacking PDG data, improving operational efficiency and supporting reservoir management decisions. Full article

Piecewise modeling of power amplifiers (PAs) typically involves assembling different polynomials to capture nonlinear behavior across different operating regions. However, recombining these sub-models can introduce discontinuities at segment boundaries, degrading prediction accuracy and potentially impacting digital predistortion (DPD) performance. This work addresses this issue by introducing a statistical framework to detect discontinuities through localized variations in the conditional mean and variance of amplitude and phase responses. Using the Vector-Switched Generalized Memory Polynomial (VS-GMP) as a case study, we propose a low-complexity post-processing smoothing technique based on a raised cosine weighting function applied at model transition regions. Unlike structural approaches, the method requires no retraining and integrates seamlessly into existing workflows as a post-processing tool. Experimental validation across two PA architectures (Doherty and Single-Stage) and multiple 5G/LTE signals (20–200 MHz bandwidth, up to 11 dB PAPR, including carrier aggregation) demonstrates consistent improvements: up to a 3 dB NMSE reduction and notable spectral error suppression. Full article

With the widespread deployment of mobile imaging sensors and smart devices, the risk of image privacy leakage is increasing daily. Protecting sensitive information in captured images has become increasingly critical. Existing image privacy protection measures usually rely on manual blurring and occlusion, which are inefficient, prone to omitting privacy information, and have an irreversible impact on the usability and quality of images. To address these challenges, this paper proposes TSLEPS (Two-Stage Localization and Erasure method for Privacy protection in Sensor-captured images). TSLEPS adopts a two-stage framework comprising a privacy target detection sub-model and a privacy text erasure sub-model. This method can accurately locate and erase the private text areas in images while maintaining the visual integrity of the images. In the stage of detecting privacy targets, an inverted residual attention mechanism is designed and combined with a generalized efficient aggregation layer network, significantly improving privacy target detection accuracy. In the stage of privacy text erasure, a texture-enhanced feature attention mechanism is proposed with an adversarial generative network for the erasure task to achieve efficient erasure of privacy texts. Moreover, we introduce the half-instance normalization block to reduce the computational load and inference time so that it can be deployed on resource-constrained mobile devices. Extensive experiments on multiple public real-world privacy datasets demonstrate outstanding performance, with privacy target detection achieving 97.5% accuracy and 96.4% recall, while privacy text erasure reaches 38.2140 dB PSNR and 0.9607 SSIM. TSLEPS not only effectively solves the privacy protection challenges in sensor-captured images through its two-stage framework, but also achieves breakthrough improvements in detection accuracy, erasure quality, and computational efficiency for resource-constrained devices. Full article

Studying spatial distribution patterns and intraspecific and interspecific associations of tree species is crucial for understanding the maintenance of biodiversity and offering insights into community dynamics and stability. The Shennongjia National Park, located in the transition zone between the (sub)tropics and the temperate climate, holds great significance for understanding how species interact with each other and coexist within forest communities. We used data from a fully mapped 25 ha montane deciduous broad-leaved forest dynamic plot at Shennongjia (SNJ) National Park, central China, to conduct a community-level evaluation of spatial distribution patterns and intraspecific and interspecific associations. We analyzed the spatial distribution patterns of 20 dominant species with univariate and bivariate g(r) functions, as well as intraspecific and interspecific associations across different life-history stages. We assessed the relative contributions of underlying processes in community assembly with three models: complete spatial randomness (CSR), heterogeneous Poisson (HP), and antecedent condition (AC). The results showed that all 20 tree species exhibited aggregated distribution patterns within a 100 m scale. After excluding the influence of environmental heterogeneity, the degree of aggregation decreased, and with the increasing spatial scale from 0 to 100 m, the distribution gradually shifted from aggregated to random or uniform appearance. Positive associations were common in different life-history stages. Negative associations were common across different species, while most of the intraspecific and interspecific associations turned out to be irrelevant when environmental heterogeneity was excluded. We concluded that habitat heterogeneity and dispersal limitation may primarily determine the spatial distribution of species in subtropical montane deciduous broad-leaved forests. This indicates that species distribution may align with environmental patterns, and interspecific correlations may exist. However, the exact responses of these species to environmental changes remain uncertain. Upcoming management approaches ought to concentrate on ongoing observation, which is crucial for mitigating how climate change might affect species distribution and community interactions, thus guaranteeing enduring stability and the conservation of biodiversity. Full article

Earthquakes, as serious natural disasters, have greatly harmed human beings. In recent years, the combination of acoustic emission technology and information entropy has shown good prospects in earthquake prediction. In this paper, we study the application of acoustic emission b-values and information entropy in earthquake prediction in China and analyze their changing characteristics and roles. The acoustic emission b-value is based on the Gutenberg–Richter law, which quantifies the relationship between magnitude and occurrence frequency. Lower b-values are usually associated with higher earthquake risks. Meanwhile, information entropy is used to quantify the uncertainty of the system, which can reflect the distribution characteristics of seismic events and their dynamic changes. In this study, acoustic emission data from several stations around the 2008 Wenchuan 8.0 earthquake are selected for analysis. By calculating the acoustic emission b-value and information entropy, the following is found: (1) Both the b-value and information entropy show obvious changes before the main earthquake: during the seismic phase, the acoustic emission b-value decreases significantly, and the information entropy also shows obvious decreasing entropy changes. The b-values of stations AXI and DFU continue to decrease in the 40 days before the earthquake, while the b-values of stations JYA and JMG begin to decrease significantly in the 17 days or so before the earthquake. The information entropy changes in the JJS and YZP stations are relatively obvious, especially for the YZP station, which shows stronger aggregation characteristics of seismic activity. This phenomenon indicates that the regional underground structure is in an extremely unstable state. (2) The stress evolution process of the rock mass is divided into three stages: in the first stage, the rock mass enters a sub-stabilized state about 40 days before the main earthquake; in the second stage, the rupture of the cracks changes from a disordered state to an ordered state, which occurs about 10 days before the earthquake; and in the third stage, the impending destabilization of the entire subsurface structure is predicted, which occurs in a short period before the earthquake. In summary, the combined analysis of the acoustic emission b-value and information entropy provides a novel dual-parameter synergy framework for earthquake monitoring and early warning, enhancing precursor recognition through the coupling of stress evolution and system disorder dynamics. Full article

Analyzing the survival and development environment, internal dynamics, and development direction of specialized farmers’ cooperatives in underdeveloped areas to enhance the vitality of the development of the agricultural industry is a key strategy for the work of the “Three Rural Issues” in China. Based on the data of 3194 specialized farmers’ cooperatives in Yunnan Province from 2000 to 2023, this paper utilizes the spatial measurement method and survival analysis method to study the spatial distribution of their survival and related influencing factors. The study found the following: (1) Cooperatives show a spatial aggregation trend from “high to high” to “low to high”, and the formation of new sub-core areas is accelerating. (2) The establishment stage of cooperatives shows an obvious annual cycle effect, and cooperatives established in the early stage show stronger survival resilience. (3) The factor of “organizational characteristics and technological innovation” significantly prolongs the survival time of cooperatives, while the factor of “establishment stage” has a negative effect. (4) The influence of a cooperative’s asset size and trademark on its operational durability tends to decrease over time, but the influence of relatedness remains relatively stable. (5) Over time, the survival and development patterns of cooperatives at the provincial level show obvious differentiation, and the clustering phenomenon of “low-high” development gradually appears in minority autonomous counties. The results of this study provide a scientific basis for deepening and strengthening the study of the basic rural business system. Full article

Tunnel disease detection and maintenance are critical tasks in urban engineering, and are essential for the safety and stability of urban transportation systems. Water stain detection presents unique challenges due to its variable morphology and scale, which leads to insufficient multiscale contextual information extraction and boundary information loss in complex environments. To address these challenges, this paper proposes a method called Deep Aggregation Network with Edge Information Supplement (DAEiS-Net) for detecting tunnel water stains. The proposed method employs a classic encoder–decoder architecture. Specifically, in the encoder part, a Deep Aggregation Module (DAM) is introduced to enhance feature representation capabilities. Additionally, a Multiscale Cross-Attention Module (MCAM) is proposed to suppress noise in the shallow features and enhance the texture information of the high-level features. Moreover, an Edge Information Supplement Module (EISM) is designed to mitigate semantic gaps across different stages of feature extraction, improving the extraction of water stain edge information. Furthermore, a Sub-Pixel Module (SPM) is proposed to fuse features at various scales, enhancing edge feature representation. Finally, we introduce the Tunnel Water Stain Dataset (TWS), specifically designed for tunnel water stain segmentation. Experimental results on the TWS dataset demonstrate that DAEiS-Net achieves state-of-the-art performance in tunnel water stain segmentation. Full article

In response to the insufficient consideration of spatio-temporal dependencies and traffic pattern similarity in traffic flow prediction methods based on federated learning, as well as the neglect of model heterogeneity and objective heterogeneity, a traffic flow prediction model based on federated learning and spatio-temporal graph neural networks is proposed. The model is divided into two stages. In the road network division stage, the traffic road network is divided into subnetworks by the dynamic time warping algorithm and the K-means algorithm, to ensure the same subnetwork has the similar traffic flow pattern. The federated learning stage is divided into two sub-stages. In the local training phase, the spatio-temporal graph neural network with an attention mechanism is utilized to create personalized models and meme models to capture the spatio-temporal dependencies of each subnetwork. At the same time, deep mutual learning is utilized to address model heterogeneity and objective heterogeneity through knowledge distillation. In the global aggregation phase, a multi-factor weighted aggregation strategy is designed to measure the contribution of each local model to the global model, to enhance the fairness of aggregation. Three sets of experiments were conducted on two real datasets, and the experimental results demonstrate that the proposed model outperforms the baseline models in three common evaluation metrics. Full article

Video-based person re-identification (ReID) aims to exploit relevant features from spatial and temporal knowledge. Widely used methods include the part- and attention-based approaches for suppressing irrelevant spatial–temporal features. However, it is still challenging to overcome inconsistencies across video frames due to occlusion and imperfect detection. These mismatches make temporal processing ineffective and create an imbalance of crucial spatial information. To address these problems, we propose the Spatiotemporal Multi-Granularity Aggregation (ST-MGA) method, which is specifically designed to accumulate relevant features with spatiotemporally consistent cues. The proposed framework consists of three main stages: extraction, which extracts spatiotemporally consistent partial information; augmentation, which augments the partial information with different granularity levels; and aggregation, which effectively aggregates the augmented spatiotemporal information. We first introduce the consistent part-attention (CPA) module, which extracts spatiotemporally consistent and well-aligned attentive parts. Sub-parts derived from CPA provide temporally consistent semantic information, solving misalignment problems in videos due to occlusion or inaccurate detection, and maximize the efficiency of aggregation through uniform partial information. To enhance the diversity of spatial and temporal cues, we introduce the Multi-Attention Part Augmentation (MA-PA) block, which incorporates fine parts at various granular levels, and the Long-/Short-term Temporal Augmentation (LS-TA) block, designed to capture both long- and short-term temporal relations. Using densely separated part cues, ST-MGA fully exploits and aggregates the spatiotemporal multi-granular patterns by comparing relations between parts and scales. In the experiments, the proposed ST-MGA renders state-of-the-art performance on several video-based ReID benchmarks (i.e., MARS, DukeMTMC-VideoReID, and LS-VID). Full article