Accurate multi-object tracking of group-housed pigs is essential for health monitoring in precision livestock farming, yet remains challenging due to similar pig appearances, frequent occlusions, dense interactions, and nonlinear motion. This paper proposes STA-PigTrack (Spatiotemporal Association Enhanced Pig Tracking), an appearance-free tracking method based on enhanced spatiotemporal association. The method introduces the Unscented Kalman Filter (UKF) to better handle nonlinear motion such as sudden accelerations and sharp turns, outperforming conventional Kalman-based models. To address identity loss from long-term occlusion and trajectory fragmentation, an AFLink module constructs a global cost matrix from spatiotemporal trajectory continuity, combined with the Hungarian algorithm to recover fragmented identities without relying on appearance embeddings. Evaluated on both a self-collected and a public dataset, STA-PigTrack achieved average HOTA, MOTA, and IDF1 scores of 83.82 %, 96.27 %, and 89.73 %, respectively, reducing identity switches by 70.2 % compared with the baseline SORT, while maintaining a real-time speed of 56 FPS. It outperformed DeepSORT, BotSORT, and OCSORT in identity consistency and robustness under heavy occlusion and high-density conditions. Using the obtained trajectories, group activity intensity and individual cumulative displacement were extracted as exploratory trajectory derived movement indicators, showing the feasibility of transforming tracking outputs into quantifiable behavioral features. STA-PigTrack provides a reliable real-time solution for individual tracking in complex farm environments and may provide a technical basis for future precision health monitoring and welfare related applications.

[Objective]Enhancing agricultural economic resilience is a critical strategic path for ensuring national food security and promoting the comprehensive implementation of rural revitalization. Against the backdrop of accelerated digital penetration in rural areas, existing research often focuses on macro-level digitalization, making it difficult to isolate the authentic contribution of digital inputs to agricultural systems. The driving effects and internal mechanisms of information and communication technology (ICT) input on agricultural economic resilience are explored in this research. Through clarifying its asymmetric impacts on resistance, recovery, and development capacities, a robust theoretical reference and empirical basis are offered for formulating differentiated digital agriculture policies that transition from traditional production modes to intelligent, resilient systems.[Methods]Based on Chinese provincial non-continuous panel data for 2012, 2015, 2017, 2018, and 2020, an advanced input-output (I-O) model framework was utilized. Leveraging the multi-regional input-output tables, the Leontief inverse matrix was employed to calculate the total consumption coefficient of the "Information Transmission, Software, and Information Technology Services" industry by the agricultural sector, which defined the total digital technology input value. Simultaneously, the entropy weight method was used to construct a comprehensive evaluation system for agricultural economic resilience. In terms of the econometric strategy, potential endogeneity was addressed by selecting the product of rural radio stations in 1988 and the previous year's Internet users as an instrumental variable (IV). The analysis was further supported by a 5% bilateral winsorization and a mediation effect model for rigorous empirical testing.[Results and Discussions]The empirical results demonstrated that digital technology input significantly enhanced overarching agricultural economic resilience. Benchmark regressions showed that the coefficient of ICT input was significantly positive at the 1% level, and the driving effect remained robust after correcting for endogeneity bias, which confirmed the core role of digital transformation in systemic risk management. Dimensional decomposition revealed a significant asymmetric characteristic: Digital technology strongly drives recovery capacity after exogenous shocks and developmental capacity during long-term evolution. However, its impact on the resistance dimension was relatively limited and exhibited a marginal negative effect. This reflected a potential technological dependence risk, where the system's increased sensitivity to power grids and network stability might weaken its original stress-resistance capacity during the onset of extreme risks. Furthermore, control variable analysis showed that per capita gross domestic product (GDP) and optimized planting structures promoted resilience, while the number of rural cooperatives exerted a negative influence, suggesting that some grassroots organizations suffered from insufficient digital adaptability. Mechanism analysis indicated that marketization, economic efficiency, and transport density were the primary transmission paths. Specifically, the marketization path contributed most significantly by reducing institutional transaction costs. Additionally, digital technology improved output efficiency through precision management and optimized transport logistics in synergy with physical infrastructure. Heterogeneity analysis showed that digital technology exhibited a clear "digital compensation" advantage in Western China, effectively offsetting natural resource endowment disadvantages.[Conclusions]This study confirms that digital input constitutes a new quality productive force that fundamentally strengthens the risk-resistance capacity of agricultural systems. The conclusions are summarized as follows: First, the empowerment of agricultural resilience by digital technology is characterized by a profound asymmetry. While it significantly improves the efficiency of systemic recovery and evolutionary development, it may simultaneously weaken original resistance due to intensified technological coupling and infrastructure dependence. Second, the reduction of institutional transaction costs through marketization is identified as the core mechanism for digital factors to exert their resilience-enhancing effects. The depth of the digital dividend is largely determined by the maturity of the market environment and its capacity for factor mobility. Third, the release of digital dividends in agriculture is heavily constrained by organizational adaptability. The lagging digital transformation and inherent structural rigidity of certain grassroots organizations have become the primary institutional bottlenecks restricting the conversion of digital technology inputs into practical systemic resilience. Ultimately, achieving a resilient agricultural economy requires a synergistic alignment between advanced digital production forces and modernized rural production relations.

Detecting keypoints in dairy cows aims to locate and track the motion trajectories of the body's joints, which plays a crucial role in behavior analysis and lameness detection. However, real farming scenarios, characterized by occlusions and large variations in object scale may result in poor detection results. Therefore, we introduce the atrous spatial pyramid pooling (ASPP) module into the shallow layers network of ResNet101, designed to improve the multi-scale feature extraction capability of the model. The ASPP module enhances the robustness of recognition for different dimensional sizes and occluded keypoints using different dilatation rates in the parallel atrous convolutional layers to expand the model's receptive field. Furthermore, seven types of motion features, including tracking up, gait symmetry, step height balance, motion speed variability, head swing amplitude, head-neck slope and back curvature are extracted simultaneously by monitoring and tracking the motion trajectory of distinct keypoints. Several of these features represent innovative extraction models and attributes, first proposed in this study. Multiple models are trained and tested on datasets containing 2,385 frames for ablation experiments. The experiments show that, in comparison with the ResNet50, MobileNet_v2_1.0, and EfficientNet-b0 backbone networks, the training error and test error of ResNet101 are reduced by 4.04-30.12 pixels and 3.81-28.14 pixels. Therefore, ResNet101 is used as the benchmark for subsequent model improvement by adding the ASPP module. The training error and test error of the ResNet101-ASPP network are reduced by 0.27 and 0.24 pixels, respectively, compared to the benchmark network. The prediction confidence improves by 1.65-2.50% at three different dairy cow object scales. In addition, the keypoints under different occlusion conditions improve considerably, especially for small-scale keypoints, demonstrating the capability of the ASPP module for multi-scale feature extraction. By analyzing the distribution of the seven features and health, mild lameness, and severe lameness in dairy cows, it is shown that all the different features play an important role in distinguishing between different levels of lameness.

Against the backdrop of global agricultural challenges such as ecological degradation and the pressure for green transformation, promoting farmers' adoption of green production behavior has become a key component in achieving sustainable agriculture. Advancing the sustainable transformation of agriculture requires contributions from farmers' social interactions. This study aims to unveil new motivational drivers behind farmers' green production behavior. By integrating the Theory of Planned Behavior with Social Learning Theory, this study developed a theoretical framework that incorporated both cognitive and social factors. Structural equation modeling and multi-group analysis were applied for empirical validation. Based on a field survey of 735 maize farmers in China, the study tested the proposed hypotheses and further analyzed indirect effects and cross-group validity. The results indicate that farmers' attitudes have the most substantial influence on intentions, with intention serving as a mediating variable indirectly affecting behavior adoption. Social learning significantly impacts all cognitive dimensions of green production behavior, validating the applicability and relevance of the proposed framework. Behavioral pathways involving social learning enhance intentions and the adoption of GPB among maize farmers. Multi-group analyses indicate heterogeneous social-learning effects by gender and education; policy should tailor technical services and peer-learning formats accordingly. Thus, improving farmers’ cognitive understanding and levels of social learning is crucial for promoting green production behavior adoption. Policymakers can facilitate green production behavior adoption by emphasizing these variables.

Cow pose estimation and real-time health monitoring are important for refined herd management, improved animal welfare, and reduced passive culling rates. However, existing multi-object pose estimation methods often struggle to adapt to multi-scale objects in complex environments and typically exhibit low accuracy in detecting occluded keypoints. To address these challenges, this study proposes a top-down deep neural network for multidairy cows pose estimation and lameness detection, which integrates lightweight object detection, multi-scale feature fusion, and comprehensive motion feature analysis to improve the robustness under complex farm conditions. First, the real-time object detector YOLOv8n is improved by introducing the Partial Convolution (PConv) and Slim-neck modules, which improve both the efficiency and accuracy of object bounding box predictions, providing a solid foundation for the subsequent pose estimation. Second, a Path Aggregation Feature Pyramid Network (PAFPN)-based multi-scale feature fusion module is introduced as the neck network within the Real-time Multi-person Pose Estimation (RTMPose). This is further supported by a transfer learning strategy to improve keypoint localization, particularly under-occlusion and scale variation conditions. The experimental results show that the improved model achieves a mean average precision (mAP) of 95.8 %, significantly outperforming the baseline model and other existing algorithms. Seven motion features, including gait symmetry, head swing amplitude, and back curvature, were extracted in real time through pose tracking and motion trajectory analysis. These features were normalized and input into a Random Forest classifier for lameness detection. The model was evaluated on a dataset of 418 dairy cows and achieved average accuracy, sensitivity, and specificity values of 93.8 %, 94.4 %, and 97.5 %, respectively. These results demonstrate that combining multiple motion features provides a more accurate assessment of lameness.

IntroductionLand use change simulation is crucial for understanding global environmental changes and guiding sustainable land management. This study conducts a bibliometric analysis of 2,147 Web of Science articles from 1988 to 2023 to summarize research trends, thematic evolutions, and future directions in land use change modeling.MethodsUsing Biblioshiny tools, the study applies quantitative analytics, co-citation network mapping, and keyword clustering.ResultsThe research reveals three developmental phases. From 1988 to 2000 (62 articles), foundational models like CLUE and CA were developed. During 2001–2016 (1,039 articles), there were advancements in coupled models and multi-scenario simulations. From 2017 to 2023 (1,046 articles), the focus shifted to integrative frameworks linking land dynamics, ecosystem services, and climate feedbacks. Annual publication outputs increased from 5 to 149, showing exponential growth. Key research themes involve computational modeling, spatiotemporal dynamics analysis, and environmental impact assessment. Recent trends highlight “river-basin,” “multi-source data fusion,” and “geographically weighted models,” indicating a move toward basin-scale simulations, machine learning integration, and policy-oriented scenarios. China, the U.S., and Germany lead in research output, with top institutions including Beijing Normal University and the Chinese Academy of Sciences. China and the U.S. have strong domestic collaborations, while European countries have higher international collaboration ratios.DiscussionThe analysis points out research gaps, such as limited integration of socio-economic drivers and insufficient cross-scale modeling. Future research should focus on developing hybrid frameworks combining process-based and data-driven models, leveraging multi-source data for accuracy, and designing scenario-based models for sustainable development goals, especially in river basins and urbanizing regions.

The bioeconomy has received significant policy attention globally, particularly in the United States and the European Union, where extensive studies have evaluated its economic importance and strategic potential. In contrast, Asia's bioeconomy, despite its substantial contributions to global biomass production and biotechnology, remains comparatively underexplored. This paper presents a study on the Chines bioeconomy value added covering the period 1995-2018, using OECD input-output statistics and the hypothetical extraction method (HEM). Our findings reveal that the Chinese bioeconomy contributes 16% to the entire economy in 2018. Furthermore, we compare the bioeconomy value added and growth rates of ten countries during the same period. The two non-OECD countries, China and India, exhibit higher percentages of bioeconomy value added, both between 15 and 19%, than the other eight OECD countries, where the percentages remain below 10%. Our results indicate that, while the total value added and bioeconomy value added fluctuate for all ten countries, the two curves follow similar trends for all countries except the United States and China. Additionally, we compare the HEM results with other methodologies and observe that the HEM and the input-based method yield similar outcomes for China, while both are considerably lower than the up- and downstream approach. This has implications for assessing the contribution of the bioeconomy for sustainable development.

With the growth of the global population and the surge in demand for animal protein, the production of animal feed has increased rapidly. This trend is expected to continue in the future. According to a report by Fortune Business Insights, the compound annual growth rate (CAGR) of global animal protein demand is projected to reach 4.49% from 2023 to 2030. However, environmental issues such as air and water pollution resulting from the feed processing industry are becoming increasingly prominent. To address these environmental challenges, intensive production methods and industrial clustering have been proposed as viable solutions. These strategies can enhance production efficiency and mitigate environmental pollution. Nonetheless, the spatial separation between upstream feed crop cultivation and downstream livestock farming, particularly in China, poses significant challenges to the development of industrial clustering. Therefore, this study investigates the spatial agglomeration patterns of China's feed processing industry, examines its geographical connection with upstream and downstream industries, and analyzes the factors driving these patterns. This study treats the feed processing industry as an independent intermediate link and offers new insights into the pollution management implications arising from the decoupling of feed crop cultivation and livestock farming. The findings reveal that, over the past two decades, China's feed processing industry has primarily concentrated in the coastal regions of East and South China, with an increasing concentration in these areas. The spatial distribution of the industry is influenced by a variety of factors. The impact of upstream raw material supply on feed processing industry is largely reflected in international imports. Downstream demand for livestock and poultry farming not only directly drives the local agglomeration of feed processing industry but also affects neighboring regions. In terms of transportation, road transport, which facilitates "door-to-door" delivery of feed, has a significantly positive impact on local agglomeration. In contrast, rail and inland river transport, due to cost limitations, suppress agglomeration in neighboring regions. Furthermore, factors such as land rent and corn temporary storage policy influence industrial clustering by affecting production costs. However, due to data availability constraints, this study does not incorporate factors such as climate and technological advancements, which warrant further exploration and refinement in future research.

Soybeans are of strategic importance to China, yet the country’s heavy reliance on imports leaves it highly exposed to policy and market disruptions. Existing studies have largely focused on the initial 2018 tariff episode, while the evolving impacts of subsequent and intensified tariff measures remain insufficiently explored. This study investigates how tariff shocks transmit through import reduction, structural reallocation, and price pass-through by employing a multi-phase difference-in-differences (DID) framework in combination with a continuous-intensity DID model. Using monthly data from January 2015 to June 2025, the analysis evaluates the effects of tariff escalation on import volumes, source-country shares, and landed import prices, thereby capturing both stage-specific dynamics and intensity-dependent responses. Robustness is verified through event-study parallel trend tests and placebo validations. The results show that (1) import volumes from the United States declined sharply and did not fully revert, indicating that tariffs disrupted long-standing trade path dependence; (2) source-country shares reallocated away from the U.S. toward South American suppliers, reinforcing diversification in China’s supply structure; and (3) tariff costs were asymmetrically passed through to prices, with U.S. soybean prices rising by approximately 43 percent, while non-U.S. prices remained relatively stable. Overall, the findings demonstrate that tariff shocks functioned as structural catalysts rather than temporary disturbances, accelerating China’s transition toward a more diversified and resilient soybean import architecture under heightened geopolitical uncertainty.

Under large-scale farming conditions, automated sow estrus detection is crucial for improving reproductive efficiency, optimizing breeding management, and reducing labor costs. Conventional estrus detection relies heavily on human expertise, a practice that introduces subjective variability and consequently diminishes both accuracy and efficiency. Failure to identify estrus promptly and pair animals effectively lowers breeding success rates and drives up overall husbandry costs. In response to the need for the automated detection of sows’ estrus states in large-scale pig farms, this study proposes a method for detecting sows’ vulvar status and estrus based on multi-dimensional feature crossing. The method adopts a dual optimization strategy: First, the Bi-directional Feature Pyramid Network—Selective Decoding Integration (BiFPN-SDI) module performs the bidirectional, weighted fusion of the backbone’s low-level texture and high-level semantic, retaining the multi-dimensional cues most relevant to vulvar morphology and producing a scale-aligned, minimally redundant feature map. Second, by embedding a Spatially Enhanced Attention Module head (SEAM-Head) channel attention mechanism into the detection head, the model further amplifies key hyperemia-related signals, while suppressing background noise, thereby enabling cooperative and more precise bounding box localization. To adapt the model for edge computing environments, Masked Generative Distillation (MGD) knowledge distillation is introduced to compress the model while maintaining the detection speed and accuracy. Based on the bounding box of the vulvar region, the aspect ratio of the target area and the red saturation features derived from a dual-threshold method in the HSV color space are used to construct a lightweight Multilayer Perceptron (MLP) classification model for estrus state determination. The network was trained on 1400 annotated samples, which were divided into training, testing, and validation sets in an 8:1:1 ratio. On-farm evaluations in commercial pig facilities show that the proposed system attains an 85% estrus detection success rate. Following lightweight optimization, inference latency fell from 24.29 ms to 18.87 ms, and the model footprint was compressed from 32.38 MB to 3.96 MB in the same machine, while maintaining a mean Average Precision (mAP) of 0.941; the accuracy penalty from model compression was kept below 1%. Moreover, the model demonstrates robust performance under complex lighting and occlusion conditions, enabling real-time processing from vulvar localization to estrus detection, and providing an efficient and reliable technical solution for automated estrus monitoring in large-scale pig farms.

With the development of precision livestock farming, in order to achieve the goal of fine management and improve the health and welfare of dairy cows, research on dairy cow motion monitoring has become particularly important. In this study, considering the problems surrounding a large amount of model parameters, the poor accuracy of multi-target tracking, and the nonlinear motion of dairy cows in dairy farming scenes, a lightweight detection model based on improved YOLO v11n was proposed and four tracking algorithms were compared. Firstly, the Ghost module was used to replace the standard convolutions in the YOLO v11n network and a more lightweight attention mechanism called ELA was replaced, which reduced the number of model parameters by 18.59%. Then, a loss function called SDIoU was used to solve the influence of different cow target sizes. With the above improvements, the improved model achieved an increase of 2.0 percentage points and 2.3 percentage points in mAP@75 and mAP@50-95, respectively. Secondly, the performance of four tracking algorithms, including ByteTrack, BoT-SORT, OC-SORT, and BoostTrack, was systematically compared. The results show that 97.02% MOTA and 89.81% HOTA could be achieved when combined with the OC-SORT tracking algorithm. Considering the demand of equipment in lightweight models, the improved object detection model in this paper reduces the number of model parameters while offering better performance. The OC-SORT tracking algorithm enables the tracking and localization of cows through video surveillance alone, creating the necessary conditions for the continuous monitoring of cows.

A comprehensive understanding of the impact of environmental regulations on agricultural green growth is crucial for formulating appropriate policy designs and achieving effective climate change mitigation. This study, based on the Porter Hypothesis framework, aims to propose our hypotheses in the influence of environmental regulations on agricultural green growth. This study first utilizes the Super-SBM model combined with the global Malmquist-Luenberger index to measure agricultural green growth in G20 countries. Subsequently, the study analyzes the impact of environmental regulations on agricultural technological innovation and green growth, highlighting variations among countries with different development levels. The results indicate: (1) Significant differences exist in agricultural green growth among G20 countries, generally exhibiting an increasing trend. (2) Environmental regulations have a significant impact on both agricultural green growth and agricultural technological innovation, but the effects of different types of environmental regulations are various. Furthermore, different conclusions are observed for countries with heterogeneous development levels. (3) There is a lag effect in the impact of environmental regulations on agricultural green growth and technological innovation, highlighting the importance of pre-establishing an environmental policy framework. (4) An inverted U-shape relationship exists between environmental regulations and agricultural green growth, while no such relationship is found in agricultural technological innovation.

With the rapid development of e-commerce in China, “live broadcast + e-commerce” has become an emerging mode of the e-commerce industry. The speedy growth of the digital economy has also provided a favorable external environment for fresh agricultural products to ride on the e-commerce express train. As a result, more and more suppliers of fresh agricultural products choose to sell their products on e-commerce platforms and conduct live streaming. This paper utilizes the e-commerce supply chain field survey data, takes tomato as the specific analyzed variety, and studies the revenue distribution pattern of each subject in the supply chain. The purpose of this paper is to explore the optimization strategy of the revenue distribution of an agricultural e-commerce supply chain, and the results show that the agricultural e-commerce supply chain effectively promotes the matching of supply and demand of medium- and high-end agricultural products, realizes the quality premium of medium- and high-end agricultural products, and significantly improves the overall revenue level of the supply chain. The field survey found that under the supply chain model of “farmers + cooperatives + e-commerce platform”, farmers, cooperatives, and e-commerce platforms gained $2865, $3098, and $1111, respectively, for operating one mu of tomatoes in a year. According to the traditional Shapley value method, from the point of view of the contribution to the cooperative income, part of the income of the farmers and cooperatives should be compensated to the e-commerce platform, and the specific results are as follows: the income of the farmers should be reduced from $2865 to $2729, the income of the cooperatives should be reduced from $3099 to $2955, and the income of the e-commerce platform should be increased from $1111 to $1390. According to the Shapley value method based on the risk coefficient correction, from the perspective of risk compensation, part of the earnings of the e-commerce platform should be compensated to farmers and cooperatives, so as to establish a more equitable and reasonable pattern of earnings distribution, with the specific results: the earnings of farmers should be increased to $3205, the earnings of cooperatives should be increased to $3148, and the earnings of the e-commerce platform should be reduced to $722.

The trade volume of agricultural products can be used to measure the global flow of virtual cropland. This paper measures the virtual cropland contained in 68 major agricultural products in China, analyzes the changes in China's virtual cropland from 2001 to 2020, and examines the main import and export agricultural product categories of virtual cropland and the main trading countries of virtual cropland. Meanwhile, according to the LMDI model, the driving factors of virtual cropland trade are categorized into five effects, and the driving factors of virtual cropland trade are discussed in depth. The main findings are as follows: the net value of virtual cropland trade is growing rapidly, the import of legume crops is the main source of the net value of virtual cropland trade, and the virtual cropland importing countries are dominated by European and American countries, and the exporting countries are dominated by Asian countries. As for the driving factors, the quality effect and the structural effect play a dampening effect on the net imports of virtual cropland, however, the dampening effect is relatively weak and cannot offset the promotion brought by the other effects. The economic effect is the main driver of virtual net imports of cropland, and the amount of trade in agricultural products also has an important role in promoting virtual imports of cropland, while the demand effect is not significant. It is suggested that China should correctly formulate reasonable cropland management policies and trade strategies, optimize the structure of virtual cropland trade, guarantee national food security and promote sustainable development.

The complex and volatile international landscape has significantly impacted global grain supply security. This study uses a complex network analysis model to examine the evolution and trends of the global major grain trade from 1990 to 2020, focusing on network topology, centrality ranking, and community structure. There are three major findings. First, the global major grain trade network has expanded in scale, with a growing emphasis on diversification and balance. During the study period, the United States, Canada, China, and Brazil were the core nodes of the network. Grain-exporting countries were mainly situated in Asia, the Americas, and Europe, and importing countries in Asia, Africa, and Europe. Second, a significant increase in the number of high centrality countries with high export capacity occurred, benefiting from natural advantages such as fertile land and favorable climates. Third, the main global grain trade network is divided into four communities, with the Americas-Europe community being the largest and most widespread. The formation of the community pattern was influenced by geographic proximity, driven by the core exporting countries. Therefore, the world needs to enhance the existing trade model, promote the multi-polarization of the grain trade network, and establish a global vision for the future community. Countries and regions should participate actively in global grain trade security governance and institutional reform, expand trade links with other countries, and optimize import and export policies to reduce trade risks.

Early lameness detection is crucial to ensure the welfare and productivity of dairy cows. However, current research on early lameness identification using wearable analysis relies on the limited robustness of indirect behavioral measures, which are susceptible to individual variations and imbalances in lameness samples. In this study, we propose a semi-supervised Long short-term memory (LSTM)-Autoencoder algorithm for early lameness detection in dairy cows through time series data reconstruction. We collected gait data from all four limbs of 30 dairy cows using four IMUs. A LSTM-Autoencoder with three LSTM hidden layers was trained to learn the time series features of healthy gaits. Each gait was reconstructed, and anomaly gaits exceeding a threshold were identified by comparing reconstructed gaits with actual gaits. The gait symmetry was measured by comparing the percentage of anomaly gait between opposite limbs as an indicator of lameness severity. With a high accuracy of 97.78% and a true negative rate of 98.33%, our integrated approach outperforms traditional methods in early lameness detection and lame limb identification, enabling real-time monitoring and timely identification of lameness. The study is the first attempt at using a time series anomaly detection framework with deep learningbased gait reconstruction for lameness detection. Wearable gait analysis offers portability and real-time capabilities, providing continuous, accurate, and comprehensive gait information unaffected by lighting and field-ofview limitations. This approach holds promise for enhancing animal welfare and optimizing management practices in the dairy industry through timely identification and continuous monitoring of lameness.

中国大豆进口量持续保持高位,其价格波动一直是关注焦点.为研究不同大豆市场价格传导的差异,选取中国、巴西、阿根廷和美国大豆市场2014年1月-2021年5月的日度价格数据,采用小波模型与Copula函数对中国与各国大豆市场价格关联进行实证分析.结果表明:中国大豆市场价格与各国大豆市场价格之间均存在显著正向关联,关联程度随时间尺度的扩大而增强,其中与巴西关联性最为显著;尾部价格联动效应存在差异,价格波动超过1个月,与阿根廷、美国分别对价格下降、上涨更敏感;中国与巴西、美国大豆市场价格都存在明显的非对称价格传导,不同国家正/负向传导效应在时间尺度上存在差异.基于研究结果,提出了提高国内油料生产,丰富进口来源,加强市场监测预警等政策建议.

目的:农产品冷链物流连接着生产与消费,是现代农业的重要特征.方法:整理了近年来我国农产品冷链物流发展情况,讨论了农产品冷链物流发展过程中存在的问题,结合我国农业基本特征分析了问题背后原因.结果:我国农产品冷链物流持续较快发展受到行业内生因素和外部环境影响,是消费需求拉动、产业技术带动和政府政策推动共同作用的结果.结论:目前,农产品冷链物流供需间存在较大缺口,技术水平相对较低.冷链物流采用成本较高,加之小农生产、产品普遍同质化,影响到相关技术采用.生产供给的季节性、地域性与冷链物流设施专用性的矛盾以及建设发展中不协调、标准不统一等问题,也影响了利用效率.针对相关问题,提出鼓励多方参与、加大研发投入、强化顶层设计、加快标准制修订等针对性建议.

In order to analyze the application and development potential of pose estimation in animal behavior recognition and animal welfare research, this paper took the deep learning-based pose estimation method as a breakthrough, summarized the research progress and directions of animal pose estimation from the perspective of two-dimensional and three-dimensional space, and introduced the common data sets and evaluation indexes. Then integrated the research results related to animal behavior recognition based on pose estimation, focused on the algorithms of keypoint detection and behavior classification and their features. Pose estimation provides skeleton information and motion features for tasks such as action recognition and behavior analysis, and became a non-contact monitoring method for animal behavior recognition and abnormal information warning. However, due to the limitation of small training data set, the research and development of animal pose estimation was relatively slow compared to human pose estimation. Therefore, the used of cross-domain learning to further improve its performance has become an emerging tool in recent years. This review expands research ideas and research methods for researchers related to intelligent animal behavior recognition, animal welfare research, and smart farming.

Sustainable food supply is affected by high energy consumption and negative environmental effects. Regarding the national strategy of "carbon peaking and carbon neutrality targets", the decoupling between energy consumption and economic growth in China's agriculture has received significant attention. Therefore, this study first presents a descriptive analysis of the energy consumption in China's agricultural sector from 2000 to 2019, before analyzing the decoupling state between energy consumption and agricultural economic growth at the national and provincial levels using the Tapio decoupling index. Finally, the logarithmic mean divisia index method is used to decompose the decoupling driving factors. The study draws the following conclusions: (1) At the national level, the decoupling of agricultural energy consumption from economic growth fluctuates among expansive negative decoupling, expansive coupling, and weak decoupling, before stabilizing in the last state. (2) The decoupling process also differs by geographic region. Strong negative decoupling is found in North and East China, and strong decoupling lasts longer in Southwest and Northwest China. (3) The factors driving decoupling are similar at both the levels. The economic activity effect promotes the decoupling of energy consumption. The industrial structure and energy intensity effects are the two main suppressive factors, whereas the population and energy structure effects have relatively weaker impacts. Therefore, based on the empirical results, this study provides evidence for regional governments to formulate policies on the relationship between the agricultural economy and energy management from the perspective of effect driven policies.