Search Results (293)

The salinization of agricultural soils is a serious threat to farming and ecological balance in arid and semi-arid regions. Accurate estimation of soil water-soluble ions (calcium, carbonate, magnesium, and sulfate) is necessary for correct monitoring of soil salinization and sustainable land management. Hyperspectral ground-based data are valuable in soil salinization monitoring, but the acquisition cost is high, and the coverage is small. Therefore, this study proposes a two-stage deep learning framework with multispectral remote-sensing images. First, the wavelet transform is used to enhance the Transformer and extract fine-grained spectral features to reconstruct the ground-based hyperspectral data. A comparison of ground-based hyperspectral data shows that the reconstructed spectra match the measured data in the 450–998 nm range, with R² up to 0.98 and MSE = 0.31. This high similarity compensates for the low spectral resolution and weak feature expression of multispectral remote-sensing data. Subsequently, this enhanced spectral information was integrated and fed into a novel multiscale self-attentive Transformer model (MSATransformer) to invert four water-soluble ions. Compared with BPANN, MLP, and the standard Transformer model, our model remains robust across different spectra, achieving an R² of up to 0.95 and reducing the average relative error by more than 30%. Among them, for the strongly responsive ions magnesium and sulfate, R² reaches 0.92 and 0.95 (with RMSE of 0.13 and 0.29 g/kg, respectively). For the weakly responsive ions calcium and carbonate, R² stays above 0.80 (RMSE is below 0.40 g/kg). The MSATransformer framework provides a low-cost and high-accuracy solution to monitor soil salinization at large scales and supports precision farmland management. Full article

Accurate and timely forecasting of cross-regional tourist flows is essential for sustainable destination management, yet existing models struggle with sparse data, complex spatiotemporal interactions, and limited interpretability. This paper presents SE-TFF, a multi-scale tourism-flow forecasting framework that couples a Squeeze-and-Excitation (SE) network with reinforcement-driven optimization to adaptively re-weight environmental, economic, and social features. A benchmark dataset of 17.8 million records from 64 countries and 743 cities (2016–2024) is compiled from the Open Travel Data repository in github (OPTD) for training and validation. SE-TFF introduces (i) a multi-channel SE module for fine-grained feature selection under heterogeneous conditions, (ii) a Top-K attention filter to preserve salient context in highly sparse matrices, and (iii) a Double-DQN layer that dynamically balances prediction objectives. Experimental results show SE-TFF attains 56.5% MAE and 65.6% RMSE reductions over the best baseline (ARIMAX) at 20% sparsity, with 0.92 × 10³ average MAE across multi-task outputs. SHAP analysis ranks climate anomalies, tourism revenue, and employment as dominant predictors. These gains demonstrate SE-TFF’s ability to deliver real-time, interpretable forecasts for data-limited destinations. Future work will incorporate real-time social media signals and larger multimodal datasets to enhance generalizability. Full article

Accurate ship berthing behavior prediction (BBP) is essential for enabling collision warnings and support decision-making. Existing methods based on Automatic Identification System (AIS) data perform well in the task of ship trajectory prediction over long time-series and large scales, but struggle with addressing the fine-grained and highly dynamic changes in berthing scenarios. Therefore, the accuracy of BBP remains a crucial challenge. In this paper, a novel BBP method based on Light Detection and Ranging (LiDAR) data is proposed. To test its feasibility, a comprehensive dataset is established by conducting on-site collection of berthing data at Dalian Port (China) using a shore-based LiDAR system. This dataset comprises equal-interval data from 77 berthing activities involving three large ships. In order to find a straightforward architecture to provide good performance on our dataset, a cascading network model combining convolutional neural network (CNN), a bi-directional gated recurrent unit (BiGRU) and bi-directional long short-term memory (BiLSTM) are developed to serve as the baseline. Experimental results demonstrate that the baseline outperformed other commonly used prediction models and their combinations in terms of prediction accuracy. In summary, our research findings help overcome the limitations of AIS data in berthing scenarios and provide a foundation for predicting complete berthing status, therefore offering practical insights for safer, more efficient, and automated management in smart port systems. Full article

To address the challenges of low recognition accuracy and substantial model complexity in crop disease identification models operating in complex field environments, this study proposed a novel hybrid model named ConvTransNet-S, which integrates Convolutional Neural Networks (CNNs) and transformers for crop disease identification tasks. Unlike existing hybrid approaches, ConvTransNet-S uniquely introduces three key innovations: First, a Local Perception Unit (LPU) and Lightweight Multi-Head Self-Attention (LMHSA) modules were introduced to synergistically enhance the extraction of fine-grained plant disease details and model global dependency relationships, respectively. Second, an Inverted Residual Feed-Forward Network (IRFFN) was employed to optimize the feature propagation path, thereby enhancing the model’s robustness against interferences such as lighting variations and leaf occlusions. This novel combination of a LPU, LMHSA, and an IRFFN achieves a dynamic equilibrium between local texture perception and global context modeling—effectively resolving the trade-offs inherent in standalone CNNs or transformers. Finally, through a phased architecture design, efficient fusion of multi-scale disease features is achieved, which enhances feature discriminability while reducing model complexity. The experimental results indicated that ConvTransNet-S achieved a recognition accuracy of 98.85% on the PlantVillage public dataset. This model operates with only 25.14 million parameters, a computational load of 3.762 GFLOPs, and an inference time of 7.56 ms. Testing on a self-built in-field complex scene dataset comprising 10,441 images revealed that ConvTransNet-S achieved an accuracy of 88.53%, which represents improvements of 14.22%, 2.75%, and 0.34% over EfficientNetV2, Vision Transformer, and Swin Transformer, respectively. Furthermore, the ConvTransNet-S model achieved up to 14.22% higher disease recognition accuracy under complex background conditions while reducing the parameter count by 46.8%. This confirms that its unique multi-scale feature mechanism can effectively distinguish disease from background features, providing a novel technical approach for disease diagnosis in complex agricultural scenarios and demonstrating significant application value for intelligent agricultural management. Full article

Soil degradation and water scarcity pose major challenges to sustainable agriculture in semiarid regions, requiring innovative strategies to enhance water use efficiency (WUE) and soil fertility. This study assessed the effects of sewage sludge biochar (SSB) on soil properties, WUE, and common bean yield through a small-scale controlled field experiment under rainfed conditions in Northeast Brazil. Four SSB application rates (5, 10, 20, and 40 t ha⁻¹) were compared with conventional NPK fertilization, treated sewage sludge (SS), and chicken manure (CM). The application of 20 t ha⁻¹ (B20) significantly improved soil organic carbon, nitrogen content, water retention, and microbial biomass. B20 also increased WUE by 148% and grain yield by 146% relative to NPK, while maintaining safe levels of potentially toxic elements (PTE) in bean grains. Although 40 t ha⁻¹ (B40) enhanced soil fertility further, it posed a risk of PTE accumulation, reinforcing the advantage of B20 as an optimal and safe dose. These results highlight the potential of SSB to replace or complement conventional fertilizers, especially in sandy soils with limited water retention. The study supports SSB application as a sustainable soil management practice that aligns with circular economy principles, offering a viable solution for improving productivity and environmental resilience in semiarid agriculture. Full article

Jilin Province is an important ecological security barrier and major grain-producing region in northeast China, playing a crucial role in ensuring ecological security and promoting regional sustainable development. This study examines ecological resilience from three dimensions: resistance, adaptability, and resilience. Based on multi-source data from 2000 to 2020, an ecological resilience indicator system was constructed. Spatial autocorrelation and OPGD models were employed to analyze temporal and spatial evolution and the driving mechanisms. The results indicate that ER exhibits an overall spatial pattern of “high in the east, low in the west, and under pressure in the central region.” The eastern mountainous areas demonstrate high and stable resilience, while the central plains and western ecologically fragile regions exhibit weaker resilience. In terms of resistance, the eastern mountainous regions are primarily forested, with high and sustained ESV, while the western sandy edge regions primarily have low ESV, making ecosystems susceptible to disturbance. In terms of adaptability, the large-scale farmland landscapes in the central regions exhibit strong disturbance resistance, while water resource adaptability in the western ecologically fragile regions has locally improved. However, adaptability in the eastern mountainous regions is relatively low due to development impacts. In terms of resilience, the eastern core regions possess stable recovery capabilities, while the central and western regions generally exhibit lower resistance with fluctuating changes. Between 2000 and 2020, the ecological resilience Moran’s I index slightly decreased from 0.558 to 0.554, with the spatial aggregation pattern remaining largely stable. Among the driving factors, DEM remains the most stable. The influence of NDVI has weakened, while temperature (TEM) and NPP-VIIRS have become more significant. Overall, factor interactions have grown stronger, as reflected by the q-value rising from 0.507 to 0.5605. This study provides theoretical support and decision-making references for enhancing regional ecological resilience, optimizing ecological spatial layout, and promoting sustainable ecosystem management. Full article

Weeds present a pervasive challenge in agricultural fields. The integration of herbicide-resistant crops with chemical weed management offers an effective solution for sustainable weed control while reducing labor inputs, particularly in large-scale intensive farming systems. Consequently, the development of herbicide-resistant cultivars has emerged as an urgent priority. In this study, we found that the G311E mutant gene of Arabidopsis MATE (multidrug and toxic compound extrusion) family transporter DTX6, designated DTX6m, confers robust resistance to bipyridyl herbicides paraquat and diquat in rice. DTX6m-overexpression lines exhibited marked resistance to these two herbicides, tolerating diquat concentrations up to 5 g/L, which is five-fold higher than the recommended field application dosage. Agronomic assessments demonstrated that grain yields of DTX6m-overexpressing plants were statistically equivalent to those of wild-type plants. Moreover, the plants displayed beneficial phenotypic changes, such as accelerated flowering and a slight reduction in height. Seed morphometric analysis indicated that in comparison with the wild-type control, DTX6m-transgenic lines exhibited altered grain dimensions while maintaining consistent 1000-grain weight. Nutritional assays further demonstrated that DTX6m increased the levels of free amino acids in seeds, while normal protein and starch contents were retained. Collectively, these results establish that DTX6m effectively boosts rice resistance to paraquat and diquat, validating DTX6m as a candidate gene for engineering plant herbicide resistance and also implying a potential role for DTX6m in amino acid homeostasis in plants. Full article

In the context of the Internet, the personalization and diversification of customer demands present a significant challenge for research on the identification, combination, and utilization of personalized demand feature elements. A key difficulty lies in achieving real-time perception, processing, and recognition of customer needs to dynamically identify and understand personalized customer intent. To address the limitations, we propose a Personalized customization-based Neural Network (PerNN), designed to enhance the performance and accuracy of recommendation systems in large-scale and complex information environments. The PerNN model introduces a Personalized Features Layer (PF), which effectively integrates multi-dimensional information—including historical interaction data, social network relationships, and users’ temporal behavior patterns—to generate fine-grained, personalized user feature representations. This approach significantly improves the model’s ability to predict user preferences. Extensive experiments conducted on public datasets demonstrate that the PerNN model consistently outperforms existing methods, particularly regarding the accuracy and response speed of personalized recommendations. The results validate the effectiveness and superiority of the proposed model in managing complex and recommendation tasks, offering a novel and efficient solution for personalized customization scenarios. Full article

Sustainable cropland management is essential in maintaining national food security. In the black soil regions of China, which are key areas for commercial grain production, sustainable land use must be achieved urgently. To address the absence of integrated, large-scale, remote sensing-based sustainability frameworks in China’s black soil zones, we developed a comprehensive evaluation system with 13 indicators from four dimensions: the soil capacity, the natural capacity, the management level, and crop productivity. With this system and the entropy weight method, we systematically analyzed the spatiotemporal patterns of cropland sustainability in the selected black soil regions from 2010 to 2020. Additionally, a diagnostic model was applied to identify the key limiting factors constraining improvements in cropland sustainability. The results revealed that cropland sustainability in Heilongjiang Province has increased by 7% over the past decade, largely in the central and northeastern regions of the study area, with notable gains in soil capacity (+15.6%), crop productivity (+22.4%), and the management level (+4.8%). While the natural geographical characteristics show no obvious improvement in the overall score, they display significant spatial heterogeneity (with better conditions in the central/eastern regions than in the west). Sustainability increased the most in sloping dry farmland and paddy fields, followed by plain dry farmland and arid windy farmland areas. The soil organic carbon content and effective irrigation amount were the main obstacles affecting improvements in cropland sustainability in black soil regions. Promoting the implementation of technical models, strengthening investment in cropland infrastructure, and enhancing farmer engagement in black soil conservation are essential in ensuring long-term cropland sustainability. These findings provide a solid foundation for sustainable agricultural development, contributing to global food security and aligning with SDG 2 (zero hunger). Full article

Effective stored-grain insect pest detection is crucial in grain storage management to prevent economic losses and ensure food security throughout production and supply chains. Existing detection methods suffer from issues such as high labor costs, environmental interference, high equipment costs, and inconsistent performance. To address these limitations, we proposed PDA-YOLO, an improved stored-grain insect pest detection algorithm based on YOLO11n which integrates three key modules: PoolFormer_C3k2 (PF_C3k2) for efficient local feature extraction, Attention-based Intra-Scale Feature Interaction (AIFI) for enhanced global context awareness, and Dynamic Multi-scale Aware Edge (DMAE) for precise boundary detection of small targets. Trained and tested on 6200 images covering five common stored-grain insect pests (Lesser Grain Borer, Red Flour Beetle, Indian Meal Moth, Maize Weevil, and Angoumois Grain Moth), PDA-YOLO achieved an mAP@0.5 of 96.6%, mAP@0.5:0.95 of 60.4%, and F1 score of 93.5%, with a computational cost of only 6.9 G and mean detection time of 9.9 ms per image. These results demonstrate the advantages over mainstream detection algorithms, balancing accuracy, computational efficiency, and real-time performance. PDA-YOLO provides a reference for pest detection in intelligent grain storage management. Full article

Drought significantly reduces global agricultural productivity and destabilizes ecosystems. As the primary grain-producing region and a key ecological buffer zone in China, Northeast China is experiencing intensifying drought stress. However, the regional-scale characteristics of refined drought and the impact mechanisms on different types of vegetation in the Northeast are rarely investigated. In this study, we analyzed the spatial and temporal characteristics of drought over 30-, 60-, 90-, 180-, 270-, and 360-day time scales in Northeast China using the Standardized Precipitation Index (SPI) based on high-precision daily precipitation data simulated by CLM3.5 from 2008 to 2023. Additionally, we used the MODIS Normalized Difference Vegetation Index (NDVI) to elucidate the response of vegetation to drought across different land use types. The results showed that SPI-30 was the most sensitive for drought detection, and there was a clear trend of drought aggravation in the northern part of the Northeast region. The strongest correlation between vegetation and drought was found in September. A significant lag in the response of vegetation to drought was observed in May, June, July, and August, with the best correlation observed at a one-month lag. In addition, the degree of response to drought varies among different types of vegetation. Grasslands are the most sensitive to drought, while woodlands and wetlands have a weaker response. This study provides a reference for assessing the dynamics of refined climates at different spatial and temporal scales and offers actionable insights for ecosystem management in climate-sensitive agricultural regions. Full article

Rice panicle detection is a key technology for improving rice yield and agricultural management levels. Traditional manual counting methods are labor-intensive and inefficient, making them unsuitable for large-scale farmlands. This paper proposes FRPNet, a novel lightweight convolutional neural network optimized for multi-altitude rice panicle detection in UAV images. The architecture integrates three core innovations: a CSP-ScConv backbone with self-calibrating convolutions for efficient multi-scale feature extraction; a Feature Pyramid Shared Convolution (FPSC) module that replaces pooling with multi-branch dilated convolutions to preserve fine-grained spatial information; and a Dynamic Bidirectional Feature Pyramid Network (DynamicBiFPN) employing input-adaptive kernels to optimize cross-scale feature fusion. The model was trained and evaluated on the open-access Dense Rice Panicle Detection (DRPD) dataset, which comprises UAV images captured at 7 m, 12 m, and 20 m altitudes. Experimental results demonstrate that our method significantly outperforms existing advanced models, achieving an AP50 of 0.8931 and an F2 score of 0.8377 on the test set. While ensuring model accuracy, the parameters of the proposed model decreased by 42.87% and the GFLOPs by 48.95% compared to Panicle-AI. Grad-CAM visualizations reveal that FRPNet exhibits superior background noise suppression in 20 m altitude images compared to mainstream models. This work establishes an accuracy-efficiency balanced solution for UAV-based field phenotyping. Full article

Phytoremediation using Avena sativa offers a sustainable strategy for mitigating sulfentrazone contamination while integrating bioenergy production. This study proposes an analysis of the bioenergy potential and the microbial metagenomic profile associated with Avena sativa in the presence and absence of sulfentrazone, aiming at the synergistic bioprospecting of microbial communities capable of biodegradation and remediation of contaminated environments. Using a randomized block design, we evaluated the bioenergy potential and rhizospheric microbial dynamics of A. sativa in soils with and without sulfentrazone (600 g ha⁻¹). Herbicide residues were quantified via UHPLC-MS/MS, and metagenomic profiles were obtained through 16S rRNA gene and ITS region sequencing to assess shifts in rhizospheric microbiota. Microbial diversity was analyzed using the Shannon and Gini–Simpson Indices, complemented by Principal Component Analysis (PCA). Bioenergy yields (biogas and ethanol) were estimated based on plant biomass. Over 80 days, the cultivation of A. sativa promoted a 19.7% dissipation of sulfentrazone, associated with rhizospheric enrichment of plant growth-promoting taxa (Bradyrhizobium, Rhodococcus, and Trichoderma), which increased by 68% compared to uncontaminated soils. Contaminated soils exhibited reduced microbial diversity (Gini–Simpson Index = 0.7), with a predominance of Actinobacteria and Ascomycota, suggesting adaptive specialization. Despite herbicide-induced stress (39.3% reduction in plant height and 60% reduction in grain yield), the biomass demonstrated considerable bioenergy potential: 340.6 m³ ha⁻¹ of biogas and 284.4 L ha⁻¹ of ethanol. The findings highlight the dual role of A. sativa in soil rehabilitation and renewable energy systems, supported by plant–microbe synergies. Scalability challenges and regulatory gaps in ecotoxicological assessments were identified, reinforcing the need to optimize microbial consortia and implement region-specific management strategies. These results support the integration of phytoremediation into circular bioeconomy models, balancing ecological recovery with agricultural productivity. Future research should focus on microbial genetic pathways, field-scale validation, and the development of regulatory frameworks to advance this green technology in global soil remediation efforts. Full article

The collapsibility of loess in the northwest region poses a significant threat to infrastructure stability. Current research predominantly separates macroscopic mechanical behaviour from mesoscopic structural characteristics, lacking a systematic methodology to quantify their interdependence. This study integrates consolidation tests, laser particle size analysis, mercury intrusion porosimetry (MIP) tests, and fractal theory to propose a multi-scale evaluation framework for assessing the structural potential of collapsible loess in strength, with on-site verification conducted. This framework quantitatively links grain size, pore potential, and connection potential to the collapsibility of loess. The experimental results indicate that loess’s high compressibility and collapsibility are primarily governed by grain size and pore potential. In contrast, the connection potential of soluble salts mitigates structural instability through ionic bonding. Field verification demonstrates a strong correlation between the three structural potentials and the subsidence coefficient (R² = 0.92, p < 0.01), validating the framework’s effectiveness in evaluating structural stability. A ternary evaluation system has been established based on fractal dimension, void ratio, and soluble salt content. These research findings provide predictive tools for managing collapse risks in loess-related projects, enabling fixed-point design of loess foundations and early warning of collapse risks. Full article

The exotic pet trade is a major driver of alien species invasions. Improper introductions or a lack of management can result in severe ecological consequences. Therefore, accurate identification of exotic pets is essential for the prevention and early warning of species invasions. This paper proposes a novel recognition method for fine-grained images of small-scale exotic pet snakes in complex backgrounds based on an improved SimCLR framework. A hierarchical window attention mechanism is introduced into the encoder network to enhance feature extraction. In the loss function, a supervised contrastive mechanism is introduced to exclude false negative samples using label information, which helps reduce representation noise and enhance training stability. The training strategy incorporates random erasing and random grayscale data augmentation techniques to improve performance further. The projection head is constructed using a two-layer multilayer perceptron (MLP), and the cosine annealing schedule combined with the AdamW optimizer is adopted for learning rate adjustment. Experimental results on a self-constructed dataset demonstrate that the proposed model achieves a recognition accuracy of 97.5%, outperforming existing baseline models. This study fills a gap in exotic pet snake classification and provides a practical tool for species invasion prevention and early detection. Full article