Search Results (202)

Laser beam scanning (LBS) projection systems based on MEMS micromirrors offer advantages such as compact size, low power consumption, and vivid color performance, making them well suited for applications like AR glasses and portable projectors. Among various scanning methods, raster scanning is widely adopted; however, it suffers from artifacts such as dark bands between adjacent scanning lines and non-uniform distribution of the scanning trajectory relative to the original image. These issues degrade the overall viewing experience. In this study, we address these problems by introducing random variations to the slow-axis driving signal to alter the vertical offset of the scanning trajectories between different scan cycles. The variation is defined as an integer multiple of 1/8 of the fast-axis scanning period ($1/{\mathrm{f}}_{\mathrm{h}}$) Due to the temporal integration effect of human vision, trajectories from different cycles overlap, thereby enhancing the scanning fill factor relative to the target image area. The simulation and experimental results demonstrate that the maximum ratio of non-uniform line spacing is reduced from 7:1 to 1:1, and the modulation of the scanned display image is reduced to 0.0006—below the human eye’s contrast threshold of 0.0039 under the given experimental conditions. This method effectively addresses scanning display artifacts without requiring additional hardware modifications. Full article

SA-ConvLSTM is a recently proposed spatiotemporal model for total electron content (TEC) prediction, which effectively catches long-term temporal evolution and global-scale spatial correlations in TEC. However, its reliance on standard convolution limits spatial feature extraction to fixed regular regions, reducing the flexibility for irregular TEC variations. To address this limitation, we enhance SA-ConvLSTM by incorporating deformable convolution, proposing SA-DConvLSTM. This achieves adaptive spatial feature extraction through learnable offsets in convolutional kernels. Building on this improvement, we design ED-SA-DConvLSTM, a TEC spatiotemporal prediction model based on an encoder–decoder architecture with SA-DConvLSTM as its fundamental block. Firstly, the effectiveness of the model improvement was verified through an ablation experiment. Subsequently, a comprehensive quantitative comparison was conducted between ED-SA-DConvLSTM and baseline models (C1PG, ConvLSTM, and ConvGRU) in the region of 12.5° S–87.5° N and 25° E–180° E. The experimental results showed that the ED-SA-DConvLSTM exhibited superior performance compared to C1PG, ConvGRU, and ConvLSTM, with prediction accuracy improvements of 10.27%, 7.65%, and 7.16% during high solar activity and 11.46%, 4.75%, and 4.06% during low solar activity, respectively. To further evaluate model performance under extreme conditions, we tested the ED-SA-DConvLSTM during four geomagnetic storms. The results showed that the proportion of its superiority over the baseline models exceeded 58%. Full article

The saddle’s adaptability to the rider–horse pair’s biomechanics is essential for equestrian comfort and performance. However, approaches to dynamic evaluation of saddle fitting are still limited in equestrian conditions. The purpose of this study is to propose a method of quantifying saddle adaptation to the rider–horse pair in motion. Eight rider–horse pairs were tested using four similar saddles with small modifications (seat depth, flap width, and front panel thickness). Seven inertial sensors were attached to the riders and horses to measure the active range of motion of the horses’ forelimbs and hindlimbs, stride duration, active range of motion of the rider’s pelvis, and rider–horse interaction. The results reveal that even small saddle changes affect the pair’s biomechanics. Some saddle configurations limit the limbs’ active range of motion, lengthen strides, or modify the rider’s pelvic motion. The temporal offset between the movements of the horse and the rider changes depending on the saddle modifications. These findings support the effect of fine saddle changes on the locomotion and synchronization of the rider–horse pair. The use of inertial sensors can be a potential way for quantifying the influence of dynamic saddle fitting and optimizing saddle adaptability in stable conditions with saddle fitter constraints. Full article

With the rapid development of intelligent transportation, accurately predicting vehicle trajectories is crucial for ensuring road safety and enhancing traffic efficiency. This paper proposes a trajectory prediction model that integrates a diffusion model framework with trajectory features of target and neighboring vehicles, as well as driving intentions. The model uses historical trajectories of the target and adjacent vehicles as input, employs Long Short-Term Memory (LSTM) networks to extract temporal features, and dynamically captures the interaction between the target and neighboring vehicles through a multi-head attention mechanism. An intention module regulates lateral offsets, and the diffusion framework selects the most probable trajectory from various possible predictions, thereby improving the model’s ability to handle complex scenarios. Experiments conducted on real traffic data demonstrate that the proposed method outperforms several representative models in terms of Average Displacement Error (ADE) and Final Displacement Error (FDE), without sacrificing efficiency. Notably, it exhibits higher robustness and predictive accuracy in high-interaction and uncertain scenarios, such as lane changes and overtaking. To the best of our knowledge, this is the first application of the diffusion framework in vehicle trajectory prediction. This study provides an efficient solution for vehicle trajectory prediction tasks. The average ADE within 1 to 5 s reached 0.199 m, while the average FDE within 1 to 5 s reached 0.437 m. Full article

Land cover changes have significant implications for ecosystem services, influencing agricultural productivity, soil stability, hydrological processes, and biodiversity. This study assesses the impacts of land use and land cover (LULC) change on soil erosion in the Upper Guder River catchment, Ethiopia, from 1986 to 2020. We analyzed Landsat imagery for three periods (1986, 2002, and 2020), achieving a classification accuracy of 89.21% and a kappa coefficient of 0.839. Using the Revised Universal Soil Loss Equation (RUSLE) model, we quantified spatial and temporal variations in soil erosion. Over the study period, cultivated land expanded from 51.89% to 78.40%, primarily at the expense of shrubland and grassland, which declined to 6.61% and 2.98%, respectively. Forest cover showed a modest decline, from 13.60% to 11.24%, suggesting a partial offset by reforestation efforts. Built-up areas nearly tripled, reflecting increasing anthropogenic pressure. Mean annual soil loss increased markedly from 107.63 to 172.85 t ha⁻¹ yr⁻¹, with cultivated land exhibiting the highest erosion rates (199.5 t ha⁻¹ yr⁻¹ in 2020). Severe erosion (>50 t ha⁻¹ yr⁻¹) was concentrated on steep slopes under intensive cultivation. These findings emphasize the urgent need for integrated land management strategies that stabilize erosion-prone landscapes while improving agricultural productivity and ecological resilience. Full article

Chirped radio-frequency signals are essential waveforms in radar systems. To enhance resolution and improve the signal-to-noise ratio through higher energy transmission, chirps with high time–bandwidth products are highly desirable. Photonic technologies, with their ability to handle broad electrical bandwidths, have been widely employed in the generation, filtering, processing, and detection of broadband electrical waveforms. In this work, we propose a photonics-based large-TBWP RF chirp generator utilizing dual optical frequency combs with a small difference in the repetition rate. By employing dispersion modules for frequency-to-time mapping, we convert the spectral interferometric patterns into a temporal RF sinusoidal carrier signal whose frequency is swept through the optical shot-to-shot delay. We derive analytical expressions to quantify the system’s performance under various design parameters, including the comb repetition rate and its offset, the second-order dispersion, the transform-limited optical pulse width, and the photodetector’s bandwidth limitations. We benchmark the expected system performance in terms of RF bandwidth, chirp duration, chirp rate, frequency step size, and TBWP. Using realistic dual-comb source parameters, we demonstrate the feasibility of generating RF chirps with a duration of 284.44 $\mathsf{\mu }$s and a bandwidth of 234.05 GHz, corresponding to a TBWP of $3.3\times {10}^7$. Full article

Despite the pivotal role of unmanned aerial vehicles (UAVs) in intelligent inspection tasks, existing video instance segmentation methods struggle with irregular deforming targets, leading to inconsistent segmentation results due to ineffective feature offset capture and temporal correlation modeling. To address this issue, we propose a hierarchical offset compensation and temporal memory update method for video instance segmentation (HT-VIS) with a high generalization ability. Firstly, a hierarchical offset compensation (HOC) module in the form of a sequential and parallel connection is designed to perform deformable offset for the same flexible target across frames, which benefits from compensating for spatial motion features at the time sequence. Next, the temporal memory update (TMU) module is developed by employing convolutional long-short-term memory (ConvLSTM) between the current and adjacent frames to establish the temporal dynamic context correlation and update the current frame feature effectively. Finally, extensive experimental results demonstrate the superiority of the proposed HDNet method when applied to the public YouTubeVIS-2019 dataset and a self-built UAV-Seg segmentation dataset. On four typical datasets (i.e., Zoo, Street, Vehicle, and Sport) extracted from YoutubeVIS-2019 according to category characteristics, the proposed HT-VIS outperforms the state-of-the-art CNN-based VIS methods CrossVIS by 3.9%, 2.0%, 0.3%, and 3.8% in average segmentation accuracy, respectively. On the self-built UAV-VIS dataset, our HT-VIS with PHOC surpasses the baseline SipMask by 2.1% and achieves the highest average segmentation accuracy of 37.4% in the CNN-based methods, demonstrating the effectiveness and robustness of our proposed framework. Full article

Seismic data processing is essential in hydrocarbon exploration, with normal moveout (NMO) correction being a pivotal step in enhancing seismic signal quality. However, conventional NMO correction often suffers from the stretch effect, which distorts seismic reflections and degrades data quality, especially in long-offset data. This study addresses the issue by analyzing synthetic models and proposing a nonhyperbolic stretch-free NMO correction technique. The proposed method significantly improves seismic data quality by preserving up to 90% of the original amplitude, maintaining frequency content stability at 30 Hz, and achieving a high reduction of stretch-related distortions. Compared to conventional NMO, our technique results in clearer seismic gathers, enhanced temporal resolution, and more accurate velocity models. These improvements have substantial implications for high-resolution subsurface imaging and precise reservoir characterization.This work offers a robust and computationally efficient solution to a longstanding limitation in seismic processing, advancing the reliability of exploration in geologically complex environments. Full article

The present study examined how making judgments of learning (JOLs) vs. not making judgments of learning (no-JOLs) influences item and sequential memory in collaborative contexts. According to the item-order hypothesis, making JOLs improves memory for specific items (i.e., item memory) but disrupts sequential memory where memory for temporal relationships between items is required. If JOLs do enhance item memory performance, the study predicts they may effectively eliminate collaborative inhibition through a compensatory enhancement mechanism. Specifically, the magnitude of JOL-induced memory improvement appears to be greater in collaborative groups than in nominal groups. This differential enhancement likely offsets the typical memory impairment caused by collaborative retrieval interference, resulting in statistically equivalent final performance between groups. Consequently, the collaborative inhibition effect may disappear under JOL conditions. This study employed a 2 (group: collaborative vs. nominal; between-subjects) × 2 (metamemory monitoring: with vs. without judgments of learning; within-subjects) × 2 (test type: recognition vs. sequential reconstruction; within-subjects) mixed factorial design. The findings indicated that making judgments of learning significantly enhanced item memory performance while having no noticeable effect on sequential memory. It suggests that the reactivity effect is only present in item memory. Additionally, it was found that both item recognition and sequential memory performance were lower in the collaborative group compared with the nominal group, highlighting the presence of collaborative inhibition. These results suggest that the reactivity effect and collaborative inhibition are two distinct memory phenomena that do not affect each other. Full article

Satellite image-based farmland classification plays an essential role in agricultural monitoring. However, typical tiling-based classification approaches, which extract patches at fixed offsets within each image during training, often suffer from structural issues such as patch duplication, limiting training diversity. Additionally, farmland classification frequently exhibits class imbalance due to uneven cultivation areas, resulting in biased training toward majority classes and poorer performance on minority classes. To overcome these issues, we propose Class-Balanced Random Patch Training, which combines Random Patch Extraction (RPE) and Class-Balanced Sampling (CBS). This method improves patch-level diversity and ensures balanced class representation during training. We evaluated our method on the FarmMap dataset, using a validation set from the same region and year as the training data, and a test set from a different year and region to simulate domain shifts. Our approach improved the F1 scores of minority classes and overall performance. Furthermore, our analysis across varying levels of class difficulty showed that the method consistently outperformed other configurations, regardless of minority-class difficulty. These results demonstrate that the proposed method offers a practical and generalizable solution for addressing class imbalance in tiling-based remote sensing classification, particularly under real-world conditions with spatial and temporal variability. Full article

This study examines the impact of economic policy uncertainty (EPU) on agricultural land conversion. Using a newspaper-based index of EPU and a comprehensive panel dataset covering 270 prefecture-level cities in China, we estimate a city fixed effects model to explore this relationship. Our results indicate that a one-standard-deviation increase in EPU leads to a 22.2% increase in the conversion of agricultural land to urban residential, commercial, and industrial uses. This finding suggests that the surge in EPU triggered by the global financial crisis accounts for approximately 45% of the increase in agricultural land conversion. The adverse effect on agricultural land preservation mainly stems from intensified fiscal pressures and heightened demands on local governments to meet economic growth targets. To address potential endogeneity concerns, we employ the one-period lagged U.S. EPU index and its temporal variations as an instrument for China’s EPU, leveraging cross-country spillover effects. Our instrumental variable estimates confirm the validity of the land conversion effect and its underlying mechanisms. Furthermore, we find that the effects of EPU are particularly pronounced in cities located in non-eastern China and those that depend heavily on fixed asset investment for local economic development. Finally, our analysis of potential policy interventions to mitigate EPU-induced agricultural land loss suggests that strengthening market-oriented reforms and reducing province-level quotas on agricultural land conversion can effectively offset the impact of rising EPU. Full article

In recent years, both anthropogenic and climate changes have caused the depletion of surface water resources, shifts in rainfall and accelerations in temperature, which indicates the importance of their examination to flood forecasting analyses. This paper studies the importance of synchronised water management strategies, considering upstream and downstream dynamics using field data from 1986 to 2018. Seasonal and decadal variations show the need for adaptive management strategies to address potential climate change impacts on discharge, precipitation and temperature patterns in the Zarrinehrud River, Iran. The regression analysis was considered via R² values, and the statistical analysis was regarded by p-values. The regression analysis of monthly river peak discharge indicates strong correlations between the discharge of specific months (September–October upstream, December–January downstream). By the 2000s and 2020s, both stations showed a shift in peak precipitation to the spring months (April–May for upstream and May–June for downstream). This confirms a synchronisation of rainfall trends, which are influenced by climate changes or regional hydrological patterns. This temporal offset between stations confirms the spatial and seasonal variation in rainfall distribution across the basin. Higher temperatures during the dominant months, particularly late summer to early autumn, accelerate snowmelt from upstream catchments. This aligns with the river discharge peaks observed in the hydrograph. The statistical analysis of river peak discharge indicated that the Weibull (p-value = 0.0901) and the Lognormal (p-value = 0.1736) distributions are the best fitted distributions for the upstream and downstream stations, respectively. Full article

Quantifying the spatio-temporal dynamics of forest biomass in both natural and planted forests over large areas has proven challenging. Using a remote sensing data-based method, this study presents a novel approach to separate the biomass gains and losses of planted forests and natural forests and to quantify their independent contributions to total forest biomass changes. Annual forest biomass data were calculated using 1 km spatial resolution maps of planted and natural forests in China for 2005–2020. Planted forest biomass increased substantially from 1.81 Pg C in 2005 to 3.11 Pg C in 2020 at a rate of 0.086 Pg C yr⁻¹. In contrast, natural forests remained relatively stable at 6.44 Pg C over the same period. Driven largely by extensive afforestation efforts, planted forests accounted for 100% of the increase in China’s forest biomass. Notably, 86.2% of the planted forest biomass and 70.3% of the natural forest biomass were located in southern China, which has a warmer climate. The area’s expansion of newly planted forests (i.e., young forests) contributed all of the total increase in biomass carbon storage (1.30 Pg C) in the planted forest category from 2005 to 2020. Forests planted before 2005 with mid-to-old tree age, together with natural forests, played a minor role in the total increase in forest biomass in China during this period. This is likely due to forest harvesting and natural disasters in these forests offsetting the growth of natural forests and mid-to-old-age planted forests over the 2005 to 2020 interval. This study highlights the complex and distinct biomass dynamics of planted and natural forests in China, which are subject to both human management and natural disturbances. Full article

The identification of leakage sources in underwater natural gas pipelines (UNGPs) remains a critical challenge due to complex environmental conditions. In this study, we propose a novel simulation–optimization method, integrating numerical bubble plume dynamics models with surrogate models to enable accurate leakage parameter inversion. First, a bubble plume underwater motion simulation model was developed based on the actual conditions of the study area to predict the future spatial and temporal variation characteristics of the bubble plumes in certain wave fields. Then, the simulation–optimization method was applied to determine the leakage velocity and offset distance of the underwater gas pipeline leakage source via inversion. To reduce the computational load of the optimization model by repeatedly invoking the simulation model, the Kriging method and a backpropagation (BP) neural network were used to build surrogate models for the numerical model. Finally, the optimized surrogate model was solved using the simulated annealing method, and the inverse identification results were obtained. The experimental results show that both methods can achieve a high inversion accuracy. The relative error of the Kriging model is no more than 12%, and the running time is 13 min. Meanwhile, based on the BP neural network surrogate model, the relative error of the BP neural network model is about 14%, and the running time is 2.5 min. Full article

Coastal blue carbon ecosystems serve as vital carbon sinks in global climate regulation, yet their long-term carbon storage dynamics remain poorly quantified at regional scales. This study quantified the spatiotemporal evolution of mangrove and salt marsh carbon storage in Guangdong Province, China, over three decades (1986–2020), by integrating a new mangrove and salt marsh detection framework based on Landsat image time series and the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model. The proposed detection framework provided two coastal vegetation detection methods, exploring the potential of utilizing phenological features to improve the mangrove and salt marsh discrimination accuracy with Landsat data. The overall accuracies of both mangrove and salt marsh detection results exceeded 90%, suggesting good consistency with the validation data. The mangrove extent showed a trend of decreasing from 1986 to 1995, then fluctuated from 1995 to 2005, and presented an upward trend from 2005 to 2020. The overall trend of the salt marsh area was upward, with small fluctuations. The mangrove carbon storage in Guangdong increased from 414.66 × 10⁴ Mg C to 490.49 × 10⁴ Mg C during 1986–2020, with Zhanjiang having the largest mangrove carbon storage increase. The salt marsh carbon storage in Guangdong grew from 8.73 × 10⁴ Mg C in 1986 to 14.39 × 10⁴ Mg C in 2020, with Zhuhai as the salt marsh carbon sequestration hotspot. The temporal dynamics of carbon storage in mangroves and salt marshes could be divided into three stages, namely a decreasing period, a fluctuating period, and a rapid increase period, during which ecological and economic policies played a crucial role. The multi-decadal blue carbon datasets and their temporal-spatial change analysis results here can provide a scientific basis for nature-based climate solutions and decision-support tools for carbon offset potential realization and sustainable coastal zone management. Full article