Cattle behavior recognition based on feature fusion under a dual attention mechanism

In recent years, artificial intelligence has been widely used in such fields as agricultural informatization, precision agriculture and precision animal husbandry. Due to limited research on deep learning in real-time agricultural and pastoral situations, deep learning and computer vision have become very important topics in the agricultural field. Recent studies have shown that the fusion of features under different attention mechanisms will help advance the utilization of such features, and will thus influence the accuracy and generalization ability of the models used. In this paper, we propose a lightweight network structure based on feature fusion under a dual attention mechanism with the same activation and joint loss functions. More specifically, we propose an innovative method to improve the network structure of two different attention mechanisms, and achieve feature fusion by combining the two. At the same time, we keep the activation functions consistent with those of the original network structure, and we develop a joint loss function to expand the use of various features. We also take the novel approach of applying the trajectory behavior analysis method to walking and standing. Experiments using both a publicly available data set and a data set obtained from a farm show that our algorithm achieves state-of-the-art performance in terms of accuracy and generalization ability, as compared to other methods.     

Global sensitivity analysis of yield output from the water productivity model

This study includes a global sensitivity analysis of the water productivity model AquaCrop. The study rationale consisted in a comprehensive evaluation of the model and the formulation of guidelines for model simplification and efficient calibration. The global analysis comprehended a Morris screening followed by a variance-based Extended Fourier Amplitude Sensitivity Test (EFAST) under diverse environmental conditions for maize, winter wheat and rice. The analysis involved twenty-two different climate-crop-soil-meteorology combinations. The main objectives were to distinguish the model's influential and non-influential parameters, and to examine the yield output sensitivity. For the AquaCrop model, a number of non-influential parameters could be identified. Making these parameters fixed would be a step towards model simplification. Also, a list of influential parameters was identified. Despite the dependence of parameter ranking on environmental conditions, guiding principles for priority parameters were formulated for calibration in diverse conditions, valuable to model users. For this model that focuses on modelling yield response to water, parameters describing crop responses to water stress were not often among those showing highest sensitivity. Instead, particular root and soil parameters, relevant in the determination of water availability, were influential under various conditions and merit attention during calibration. The considerations made in this study about sensitivity analysis method (Morris vs. EFAST), prior parameter ranges, target functions and ranking variation according to environmental conditions can be extrapolated to other conditions and models, if done with the necessary precaution.     

Strengths and limitations on the operating mechanisms of traceability system in agro food,China

Traceability system can effectively trace food quality and reduce safety scares. Recent researches on traceability systems in China focus on technology innovation, traceability system management, and determinants of traceability system implementation. This paper proposes four criteria to analyze strengths and limitations of the operating mechanisms. The result shows the operating mechanism of traceability system in Chinese agribusiness can be classified into three categories and none of operating mechanism of traceability system completely meets the agribusiness traceability service requirement, and each of the three operating mechanisms has its own strengths and limitations. It is suggested that an integrated mechanism is needed to implement traceability system in agribusiness.     

农业害虫检测的深度学习算法综述

害虫检测是害虫测报的关键步骤,对于害虫防治具有重要意义,也是保证农作物产量和品质的前提。近年来,随着卷积神经网络的迅速发展,害虫检测技术进入智能化时代,使用深度学习相关技术实现精确的害虫检测已成为研究人员重点关注的课题。为了促进深度学习害虫检测技术的发展,对检测算法和现有数据集进行综述。总结了当前面临的数据匮乏、小目标检测、多尺度检测和密集与遮挡检测等四大难点问题,并分析了其主要成因。重点针对以上难点问题,总结归纳了近年来提出的深度学习害虫检测算法的改进策略和技术细节,以及面向实际场景的应用算法,对比分析了各类算法的性能表现、改进策略的适用场景及其优缺点。从面向复杂检测场景、解决数据匮乏问题、模型增量更新和应用落地等方面分析并展望了未来的研究趋势。     

基于BP和GRNN神经网络的冬小麦冠层叶绿素高光谱反演建模研究

根据高光谱遥感获得的冬小麦冠层数据,把由逐步回归方法和基于遗传算法(GA)的广义回归神经网络(GRNN)筛选到的光谱参数作为网络输入,冠层叶绿素含量作为网络输出,采用线性逐步回归方法、反向传播神经网络(BPNN)和GRNN来构建反演模型,模拟结果表明,GRNN和BPNN的预测精度要高于逐步回归方法,其RMSE分别为0.36 mg/g、0.52 mg/g和0.98 mg/g。由于GRNN可应用于小样本问题的学习,比BPNN对叶绿素具有更好的预测和泛化能力。     

New principles in metallomesogen structure–magnetism correlations

Whereas binuclear salicylaldimine complexes and their mononuclear parents are waxlike solids with long chain N-substituents, they are liquid crystalline (mesogenic) when such substituent chains are placed on both the ligand rings and the N-atoms. The structure of the non-mesogenic binuclears is determined via X-ray crystallography, and via the strength of magnetic coupling for the mesogenic ones. In the non-mesogenic binuclears the central metal environments have a marked distortion towards tetrahedral (>36°; scale 0° for planar, 90° for tetrahedral), while in the mesogenic ones, the stronger magnetic coupling indicates that this is closer to planar (25°).     

Modeling and optimization of two-stage procurement in dual-channel supply chain

An online marketing platform should be designed to fairly take the benefits of buyers and suppliers into consideration based on their risk preferences and business strategies. In this paper, the dual-channel supply chain models are developed to implement the risk-averse strategy for buyers and risk-neutral strategy for suppliers, respectively. The buyers under the consideration are the manufacturers who acquire raw materials, parts, or components to make their final products. The major factors in the developed models include the risk preferences of buyers and suppliers, random price fluctuations of goods, and varying demands of final products. To reflect the purchasing practice of a manufacturer, (1) a supply chain is considered to have two supply channels, i.e., contract-based purchase with a lead-time before the goods are used and a direct purchase from online spot markets when the goods are used; (2) the time factor on decision making is specially taken into account, and the procurements are divided into the contract stage of purchase and online stage of purchase. Gaming analysis is conducted to develop the supply chain models for manufactures and suppliers to implement their purchasing or pricing strategies. The simulation is conducted and the result has shown that two-stage purchases in a dual-channel supply chain have improved the performances of suppliers and manufacturers in terms of the profits they can make, their supply–demand relations, and their adjustability to uncertainties in globalized and segmented markets. The proposed model has its significance for manufacturers to better control the price risk of goods and the demand risk of final products; on the other hand, suppliers can benefit from adjusting dynamic sales using online spot markets.     

果蔬类农产品多源追溯系统设计与实现

目的 针对农产品质量安全问题, 以果蔬类农产品为研究对象, 通过分析农产品生产过程, 研究农事操作、生产环境、检测信息、生产视频的信息采集技术、数据交换技术、产品溯源技术。方法 采用C/S架构农产品企业的管理系统与B/S架构的农产品追溯系统相结合的模式, 设计开发农产品多源追溯系统。结果 实现了农产品环境数据自动采集、生产过程视频、生产农事记录的信息化管理, 并能够通过手机短信、触摸屏、手机二维码扫描、追溯网站等多种追溯方式对农产品生产信息追溯, 为提高农产品质量安全追溯水平提出了可行的解决方案。结论 该系统在山东滨州的追溯服务中心和农产品生产企业进行了应用示范。     

A simple calculation model for radiant heat transfer in a paper dryer

A simple and practical method for calculating radiant heat transfer in a directly heated industrial paper dryer is presented. The paper dryer is modelled as an enclosure of three isothermal surfaces. A short overview of the theory of radiant heat transfer between N surfaces is given. The calculations show good agreement with experimental results. A numerical example of the calculation procedure and a FORTRAN-V program are included.     

A framework to use crop models for multi-objective constrained optimization of irrigation strategies

This paper discusses an innovative framework to use crop models which combines sensitivity analysis, uncertainty analysis and constrained optimisation runs for irrigation optimisation purposes, facing competing constraints on several agricultural variables (e.g. crop yield, total irrigation amount, financial expectations). For simplicity, this ex-post optimisation relies on direct calculations only, exploiting the dispersions on the target variables. The screening of the parameter space for sensitivity analysis yields a reference dispersion which is expectedly reduced by reducing the uncertainties in the sensitive parameters and/or climatic forcings. Additional dispersions are calculated to evaluate if the management controls on irrigation strategies (amounts, triggers, periods) are more influential on model predictions than the remaining uncertainties on the soil, plant, irrigation and climatic inputs, eventually allowing optimisation. As a case study, the Optirrig model is used. A discussion proposes future ways to convert diagnostics into real-time near-optimal decision rules, for example through learning algorithms.