Search Results (1,140)

Accurate pest recognition is an important foundation for intelligent plant protection, precision pesticide application, and sustainable agricultural management. However, in real field environments, pest targets are often small in scale, severely occluded, and embedded in complex backgrounds, which limits the performance of existing supervised learning methods under low-annotation and cross-scenario conditions. To address these issues, a multimodal self-supervised pretraining framework is proposed for pest recognition, in which field pest images and environmental sensor data are integrated to construct pest representations with environmental awareness. In this framework, image features, including pest morphology, leaf texture, and damaged regions, are first extracted through a visual encoding branch, while temporal variation features of ecological factors, including temperature, humidity, illumination, soil moisture, rainfall, and wind speed, are modeled through an environmental encoding branch. On this basis, a cross-modal contrastive consistency module is designed to align visual and environmental representations, a temporal consistency self-supervised module is introduced to characterize the continuous evolutionary relationship between pest occurrence and environmental changes, and a multimodal collaborative representation fusion module is constructed to adaptively integrate information from different modalities. The experimental results show that the proposed method achieves favorable performance in the pest recognition task, with Accuracy, Precision, Recall, and F1-score reaching $94.37\%$, $93.96\%$, $93.42\%$, and $93.69\%$, respectively, outperforming ConvNeXtV2-T, ViT-B/16, Swin-T, SimCLR, MAE, and the conventional Image + Sensor fusion method. The ablation experiments further show that, after removing the cross-modal contrastive consistency module, the temporal consistency self-supervised module, and the multimodal collaborative fusion module, the F1-score decreases to $91.00\%$, $91.36\%$, and $90.49\%$, respectively, thereby demonstrating the contribution of each module. This study provides a viable multimodal self-supervised learning approach for AI-driven intelligent pest recognition, early warning, and precision control in agriculture. Full article

This review provides a comprehensive synthesis of robotic fruit harvesting systems, with a particular focus on the system-level integration of perception, manipulation, and fruit detachment within autonomous harvesting environments. Recent advances in machine vision, deep learning, sensor fusion, robotic end-effectors, grasping strategies, and motion planning are critically analyzed alongside cutting, pulling, and vibration-based detachment mechanisms under unstructured orchard conditions. Beyond component-level analysis, this review emphasizes the critical role of perception–action coupling and highlights key system integration challenges, including localization errors, perception-to-action latency, and environmental variability, which continue to limit reliable field deployment. In addition, orchard and pre-harvest-related factors such as canopy structure, fruit distribution, and detachment force variability are examined in relation to their direct impact on system performance, robustness, and harvesting efficiency. Furthermore, the review extends toward system-level considerations by incorporating performance evaluation metrics, economic feasibility, and scalability constraints, which are essential for transitioning robotic harvesting systems from experimental prototypes to commercially viable solutions, including practical field deployment in distributed and multi-robot harvesting systems. Emerging technologies, including artificial intelligence, advanced sensing, digital agriculture, and energy-aware system design, are discussed as key enablers for achieving adaptive, data-driven, and scalable autonomous harvesting. The novelty of this work lies in proposing an integrated framework that explicitly links perception, manipulation, and detachment with orchard-level constraints and deployment requirements, thereby bridging the gap between algorithmic advancements and real-world implementation of autonomous fruit harvesting systems. Full article

The detection technology of crop diseases and pests is transitioning from single sensor monitoring to intelligent perception and multimodal fusion. This paper follows the PRISMA 2020 standard and systematically reviews the relevant core literature. This paper systematically summarizes the development history of spectral sensing technology and analyzes the physical mechanisms of hyperspectral and multispectral imaging in early identification of crop diseases. The focus is on the architectural evolution of deep learning models, including lightweight convolutional neural networks (CNNs), vision transformers (ViTs) with long-range dependency modeling capabilities, and the efficient computing state space model Mamba. In addition, the research progress of spatial spectral joint learning, heterogeneous data fusion, and vision-language models (VLMs) in improving system robustness and interpretability are introduced. By synthesizing the integrated applications of UAV remote sensing, Internet of Things (IoT) edge computing and intelligent robots in staple and cash crops, this paper summarizes the implementation of the integrated system of perception, decision-making and execution. To address the issues of insufficient cross-domain generalization ability and uneven allocation of computing resources in existing models, this paper provides perspectives on the future development of agricultural artificial intelligence (AI) towards foundation model-driven, edge-intelligent collaboration, and green sustainable direction, which can provide theoretical reference for engineering applications in the field of intelligent plant protection. Full article

Urban soil contamination by potentially toxic elements (PTEs) is a recognized health concern in densely populated urban environments. Through a systematic meta-analysis of 91 peer-reviewed studies (2000–2025) reporting 12,174 sampling sites in capital and core cities, we characterized regional patterns in the spatiotemporal dynamics and health risks of eight PTEs across two well-represented continental subsets (Asia, k = 18–36 per element; Europe, k = 11–23 per element) with comparative reference to the Americas, Africa, and Oceania. Given the uneven geographic distribution of qualifying primary studies, continental comparisons should be interpreted as hypothesis-generating: Asia (k = 18–36 per element) and Europe (k = 11–23 per element) provide the statistically robust core of the synthesis, while results for the Americas (k = 3–7 for several elements), Africa (k = 4–15), and Oceania (k = 2) are presented as illustrative rather than statistically representative. Pooled concentrations followed Zn (138.59) > Pb (56.97) > Cr (54.26) > Cu (47.00) > Ni (31.94) > As (8.56) > Hg (3.13) > Cd (1.23) mg·kg⁻¹. Within the well-represented Asian and European subsets, Asian cities showed the most severe enrichment of As, Cd, Cr, and Hg (I_geo > 4 in hotspots such as Kathmandu I_geo(Cd) = 7.06 and Jinan I_geo (Hg) = 5.27), whereas European centres exhibited substantial legacy Pb accumulation (pooled mean 87.69 mg·kg⁻¹). A reproducible pollution gradient was identified across functional zones: industrial > transportation ≥ residential > commercial > agricultural > urban green areas. The deterministic non-carcinogenic Hazard Index (HI = 1.49) for children in Asia exceeded the safe threshold (HI > 1), driven primarily by As and Cr exposure via incidental soil-and-dust ingestion. Monte Carlo probabilistic assessment (N = 10,000) confirmed elevated cumulative non-carcinogenic risk at the median of the exposure distribution for children in the data-rich Asian (P50 = 1.55; P(HI > 1) = 81.9%) and European (P50 = 1.28; P(HI > 1) = 69.8%) subsets, with adults in both subsets remaining well below the safety threshold (P(HI > 1) = 0.0%). Temporal analysis revealed a decoupling between economic growth and PTE accumulation in long-established cities, together with an inverse Ni–population correlation indicative of strategic resource allocation. For Asian capital and core cities, where the evidence base is strongest (k = 18–36 per element), the present synthesis supports further investigation of risk-based, child-centric soil management as a public-health priority. For European cities (k = 11–23 per element), the same direction of risk is indicated but should be confirmed in regionally focused syntheses. Policy considerations for under-represented regions should await expansion of the primary monitoring base. Full article

Although rural livelihoods act as a critical mediator between poverty alleviation policies and sustainable outcomes, the spatial heterogeneity of this interaction remains underexplored within those agrarian systems that are crucial for food production. This study examines how China’s Targeted Poverty Alleviation policies shape livelihood strategies and the livelihood diversity of rural households across different spatial contexts in Jilin Province, a main grain production region of Northeast China. Using survey data from 2306 households, this study employs multiple logistic and linear regression models. The results indicate that (1) industrial and employment policies are associated with development-oriented strategies, whereas enterprise-driven and cash transfer policies tend to reinforce asset-based or welfare-dependent livelihoods; (2) these policy effects exhibit significant spatial heterogeneity, mediated by local agricultural productivity conditions, labor endowments, and off-farm livelihood availability; and (3) industrial policies show stronger associations with agricultural livelihoods in the east, while financial policies are more effective in sustaining agricultural engagement in the capital-constrained west. Integrating the Sustainable Livelihoods Framework with a spatial lens, this study shifts the focus of policy assessment from static outcome metrics to process-oriented analysis and reveals the mechanisms underlying the spatial divergence of livelihood strategies, providing a nuanced analytical framework for assessing the impacts of PAPs across diverse agricultural contexts. Based on these findings, this study highlights that spatially differentiated, livelihood context-sensitive policies are essential for securing sustainable and long-term poverty reduction in grain production regions, offering a replicable template for policy evaluation and practical implications for achieving SDGs 1 and 2 in agrarian regions. Full article

This study investigates the potential of several sustainable agricultural by-products—including olive stones, cork, and almond shells, which are locally available in Alentejo, Portugal—as low-cost adsorbents for the removal of methylene blue (MB) from synthetic wastewater. The biomass residues were evaluated both in their raw form and after conversion into activated carbons (ACs) through chemical activation with KOH at 973 K. The produced ACs exhibited well-developed surface areas (760–1103.5 m² g⁻¹) and porous structures (0.31–0.51 cm³ g⁻¹). The adsorbents were characterised in terms of their chemical and textural properties. Raw biomass materials presented acidic surface groups, whereas the ACs presented neutral or basic groups. Batch adsorption experiments were conducted to assess the effects of adsorbent particle size, solution pH, initial MB concentration, stirring speed, contact time, and temperature on dye removal efficiency. Among all tested materials, the ACs achieved superior MB adsorption capacities, ranging from 244.2 to 317.6 mg g⁻¹, compared to the untreated biomass adsorbents, which showed capacities between 34.1 and 46.4 mg g⁻¹. The adsorption data were best described by the Langmuir isotherm model, while the kinetic data closely followed the pseudo-second-order (PSO) model. Thermodynamic analysis revealed that MB adsorption was spontaneous and endothermic; however, the relatively low enthalpy values indicated that physical interactions contributed significantly, particularly in the case of the raw biomass adsorbents. This suggests that the PSO model may also be applicable when physical adsorption is the dominant mechanism. This work demonstrates the novel use of cork, olive stone, and almond shell biomasses and their derived ACs as sustainable adsorbents, highlighting an integrated approach that simultaneously promotes efficient wastewater treatment, waste valorisation, and circular economy-driven socio-economic development. Full article

In response to increasingly complex risks and challenges and to safeguard national agricultural product supply security, this study constructs a four-dimensional evaluation index system encompassing “Resistance-Adaptation-Recovery-Innovation”. Utilizing panel data from 30 provincial-level regions in China from 2017 to 2023, and employing a comprehensive methodology including the entropy method, Dagum Gini coefficient, Markov chain, kernel density estimation, and convergence models, this research measures the resilience of China’s agricultural product supply chain and investigates its spatiotemporal evolution patterns. The findings are as follows: Firstly, the resilience level of the national agricultural product supply chain shows overall steady improvement, but regional development is uneven, presenting a pattern of eastern regions leading, central regions maintaining steady progress, and western regions catching up. Secondly, the overall resilience difference is strongly correlated with regional variability, with the most pronounced internal disparity observed in the western region. Thirdly, the evolution of resilience exhibits path dependency characterized by the coexistence of a “low-level trap” and “high-level stability”, and less developed regions demonstrate a significant “catch-up effect” towards their more developed counterparts. Based on these findings, this study proposes countermeasures such as implementing targeted policies for different regions, establishing cross-regional coordination mechanisms, strengthening dynamic monitoring and early warning systems, and promoting innovation-driven development and structural upgrading. These efforts aim not only to enhance China’s capacity to respond to risks in its agricultural product supply chain and ensure national food security, but also to provide valuable insights for other countries facing similar challenges in building resilient agricultural systems in an increasingly uncertain global environment. Full article

This paper examines whether total factor productivity (TFP) in Israeli agriculture has genuinely slowed or declined in recent years, or whether the reported trend is primarily driven by methodological choices, data limitations, and measurement error. We compare two widely used approaches to TFP measurement—those of the Bank of Israel and the U.S. Department of Agriculture (USDA)—which differ in their definitions of output, treatment of inputs, and assumptions regarding factor shares. We reconstruct and refine the underlying datasets, addressing important limitations in the existing measures, including the omission of foreign labor, inconsistencies in agricultural land measurement, and the application of non-representative input shares. Despite data improvements and methodological adjustments, both approaches yield similar qualitative conclusions. Following rapid increase in earlier decades, TFP growth in Israeli agriculture appears to have stagnated or declined since the early 2010s. A decomposition of output growth further indicates that recent production patterns have been driven primarily by greater input intensity per unit of land rather than by technological progress or efficiency gains. As a result, agricultural output has shown little or no net growth over the past decade. We discuss potential explanations for this slowdown, including climate change, the growing reliance on reclaimed and other marginal water sources, and the long-term decline in agricultural research and development (R&D) investment relative to sectoral output. Overall, the findings suggest that the productivity slowdown is real rather than an artifact of measurement and underscore the need for renewed investment in agricultural innovation and climate adaptation to sustain domestic production and strengthen food security. Full article

Rice crop residue burning (RCRB) in India constitutes a major annual environmental crisis, contributing significantly to regional air pollution, greenhouse gas emissions, and public health deterioration across the Indo-Gangetic Plain. Despite growing policy attention, a systematic, data-driven understanding of the diverse perspectives—agricultural, environmental, economic, and socio-political—expressed across multiple textual sources remains lacking. This study proposes a large language model (LLM)-driven topic modeling pipeline leveraging TopicGPT, an instruction-tuned prompting framework, to extract and evaluate high-level thematic insights from heterogeneous text corpora related to RCRB in India. Our pipeline integrates four sequential stages—topic generation, topic refinement, multi-label topic assignment with grounded evidence, and assignment correction—operated via a locally deployed LLM through the Ollama inference framework. Post-extraction, we evaluate topic quality using ten quantitative metrics encompassing embedding-based coherence, inter-topic diversity, Silhouette score, Davies–Bouldin index, Calinski–Harabasz score, and distribution entropy, among others. Results demonstrate that the proposed pipeline effectively recovers semantically coherent and diverse topic structures from multi-source text data, offering actionable insights for policymakers and researchers addressing RCRB. Full article

Humic acid (HA) production from crop straw is often limited by lignocellulosic recalcitrance and insufficient coordination among functional microorganisms. In this study, a data-driven strategy was developed to evaluate multi-fungal inoculation timing for HA recovery from pretreated straws. Three substrate platforms, namely raw wheat straw (SW), steam-exploded corn straw (SC-SE), and ammoniated steam-exploded rice straw (SR-SE-N), were comparatively evaluated across an 81-run experimental matrix. Pretreatment markedly improved lignocellulose degradation and precursor turnover, with SR-SE-N showing the best humification performance. Based on the selected substrate, a two-factor interaction (2FI) model was established to describe the effects of inoculation timing on HA yield. The model was significant for HA prediction (R² = 0.8768, adjusted R² = 0.8398, predicted R² = 0.7795). Inoculation timing strongly affected HA formation, and within the investigated timing range, the highest HA yield was obtained under simultaneous inoculation of Aspergillus niger, Phanerochaete chrysosporium, and Candida sp. Predicted and experimental HA yields were in close agreement, supporting the reliability of the model. These results indicate that favorable fungal inoculation timing is substrate-dependent and can be effectively identified through data-driven analysis within a bounded experimental range. The study provides a practical basis for improving HA biomanufacturing from pretreated agricultural residues. Full article

Drought stress is the most pervasive abiotic constraint on global crop productivity, with projected intensification under climate change threatening the yields of staple crops including wheat, rice, maize, and legumes. Conventional breeding approaches have delivered limited gains against drought tolerance, constrained by the polygenic and multifactorial nature of stress adaptation, the complexity of genotype-by-environment interactions, and the inadequacy of field-based phenotyping under variable stress conditions. Omics technologies, including genomics, transcriptomics, proteomics, metabolomics, epigenomics, and phenomics, have substantially advanced the molecular dissection of drought tolerance by enabling high-resolution characterization of stress-responsive genes, regulatory networks, adaptive proteins, and metabolic reprogramming pathways. Specific traits targeted include root system architecture and depth, osmotic adjustment capacity through proline and glycine betaine accumulation, antioxidant defense mechanisms, ABA-mediated stomatal regulation, LEA protein accumulation, epigenetic stress memory, and yield stability under water deficit. This review systematically examines omics-based strategies for drought stress mitigation across major crops, highlighting individual omics contributions, multi-omics integration frameworks, computational tools including machine learning and AI-driven predictive modelling, and translational breeding applications. Case studies in wheat, rice, maize, and legumes illustrate how omics-driven approaches accelerate precision breeding for drought resilience through marker-assisted selection, genomic selection, and CRISPR-based gene editing. Challenges including data integration complexity, high implementation costs, limited cross-species transferability, and the need for field-scale validation of microbiome-based strategies are critically addressed. Future perspectives encompassing single-cell and spatial omics, AI-driven predictive breeding, digital agriculture integration, and international data governance frameworks are discussed. By aligning with climate-smart agriculture principles, multi-omics approaches provide a robust and transformative foundation for developing drought-resilient crop cultivars suitable for water-limited production systems worldwide. Full article

Agricultural cold chain logistics is characterized by inherent challenges—product perishability, high carbon emissions, and stringent time windows—which are further exacerbated by dynamic disruptions. Existing methods suffer from slow adaptability, unstable multi-objective convergence, and severe cold-start issues. This work falls within the broad scope of biomimetics—the science of emulating nature’s time-tested strategies to solve complex engineering problems—and bio-inspired data-driven methods and their applications in engineering control, optimization, and artificial intelligence. The proposed H-MODRL framework embodies core biomimetic principles: the Genetic Algorithm (GA) mimics Darwinian natural selection and genetic inheritance, the Sparrow Search Algorithm (SSA) abstracts the cooperative foraging and anti-predation behaviors of sparrow populations in nature, and the Arrhenius-based freshness-decay model captures the biochemical kinetics governing perishable biological products. By synergistically integrating these biological evolution principles, swarm intelligence, and deep learning, the framework tackles real-world logistics complexity in a manner directly inspired by living systems. This study presents a well-organized hybrid optimization framework (H-MODRL) that couples a three-stage hybrid evolutionary mechanism, synergistically integrating heuristic warm-start, evolutionary policy guidance, and deep reinforcement learning decision-making. First, an improved genetic algorithm combined with the earliest deadline first strategy constructs a feasible initial population satisfying hard time-window constraints. Second, a large neighborhood search-enhanced chaotic sparrow search algorithm builds a high-quality elite guidance set for policy learning. Third, a physics-based multi-objective proximal policy optimization model embedded with Arrhenius equation-derived freshness-decay kinetics performs online decision-making. Experiments demonstrate that pre-computed all-pairs shortest paths and an O(1) hash-based dynamic-disruption indexing mechanism support fast online replanning. On heterogeneous simulated terrains based on real Chinese geospatial data, H-MODRL outperforms state-of-the-art algorithms across four objectives—logistics cost, carbon emissions, terminal freshness, and delivery time—while exhibiting compact, low-variance performance distributions, thereby validating its engineering robustness and practical value in complex agricultural cold chain environments. Full article

Land-use change is a major driver of climate change and carbon-cycle imbalance, yet its dual effects on carbon emissions and carbon sinks remain insufficiently examined in the context of rural spatial restructuring. As a key form of county-level rural land-use transformation in China, the Village Merger and Resettlement (VMR) policy reshapes the spatial distribution of rural population and land by concentrating residents, vacating idle homesteads, and optimizing land allocation. Treating VMR as a quasi-natural experiment, this study uses county-level panel data and a difference-in-differences approach to evaluate its impacts on regional carbon emissions and carbon-sink capacity. The findings indicate that VMR significantly reduces regional carbon emissions while enhancing carbon-sink capacity. Mechanism analysis shows that the emission-reduction effect operates mainly through population concentration, centralized infrastructure provision, and reduced household energy consumption, whereas the carbon-sink effect is driven by land intensification, agricultural mechanization, cultivated land-use adjustment, and improved vegetation cover. Heterogeneity analysis further reveals stronger emission-reduction effects in areas with a larger pre-policy urban–rural income gap and higher urbanization rates, and stronger carbon-sink effects in areas with greater terrain fragmentation and higher elevations. These findings suggest that rural spatial restructuring can contribute to regional carbon governance. Full article

The utilization of multi-source sensing data to achieve intelligent perception and refined management of farmland has become a vital research direction in modern agriculture. However, traditional inspection approaches based solely on visual information are highly susceptible to illumination variations, occlusion, and background interference, which makes stable pest detection and accurate crop growth assessment difficult to achieve. To address these problems, we propose a multimodal target perception network for intelligent farmland inspection. By integrating UAV imagery, ground environmental sensor data, and spatial location information, joint perception of farmland pests, diseases, and crop growth status is achieved. In the proposed framework, cross-modal alignment and collaborative encoding mechanisms, a multi-scale target perception structure, and a dynamic multimodal fusion strategy are introduced to collaboratively model information within a unified semantic space. Experimental results on a constructed multimodal farmland dataset demonstrate that the proposed method achieved 87.53% Precision and 89.16% mAP in the pest and disease detection task, and 88.04% Accuracy in the crop growth assessment task, significantly outperforming several mainstream visual detection models and multimodal fusion approaches. The results indicate that this intelligent perception framework can significantly improve the robustness of farmland inspection systems, providing an effective technical pathway for AI-driven precision agriculture decision-making. This technology breaks the barrier between production-side sensing data and e-commerce demand, providing a practical technical solution for agricultural production-marketing synergy, quality premium realization and digital rural revitalization. Full article

Accurate quantification of soil organic matter (SOM) and carbon content is critical for understanding climate change, evaluating soil health, supporting agricultural sustainability, and implementing carbon sequestration policies. For decades, classical analytical and statistical approaches have underpinned soil carbon assessment, but the emergence of machine learning (ML) techniques offers new opportunities to improve prediction accuracy, scalability, and efficiency. This review summarises the current knowledge on classical and ML-based approaches for analysing SOM and carbon content. We examine the strengths, limitations, and practical applications of conventional methods, including wet chemistry, dry combustion analysis, and geostatistical techniques, alongside modern ML approaches such as random forests (RFs), gradient boosting machines, neural networks, deep learning, and hybrid ML-geostatistical frameworks. Special emphasis is placed on comparative analysis across dimensions, including prediction accuracy, computational requirements, data availability needs, interpretability, uncertainty quantification, and scalability. Soil carbon stocks and dynamics are tightly regulated by indigenous soil microbial communities and their management-driven alterations, creating substantial biologically driven variation that remains difficult to capture with current modelling approaches. We therefore explore hybrid approaches that integrate classical pedological knowledge with ML capabilities. Finally, we discuss emerging challenges, future research directions, and the complementary role these approaches play in advancing soil carbon science. This review concludes that neither classical nor ML approaches alone are sufficient for accurate carbon assessment across diverse scales and environments. Instead, their strategic integration, combining classical mechanistic grounding alongside machine learning’s scalability, represents the most promising path toward realistic soil carbon evaluation for climate change mitigation and agricultural sustainability. Full article