改进ShuffleNet V2的轻量级农作物病害识别方法

针对目前有关深度学习的农作物病害识别方法中存在模型较为复杂,部署在计算资源有限的边缘设备和移动终端上适应性不强,实时准确识别作物病害较差的问题,提出一种改进ShuffleNet V2的轻量级农作物病害识别方法。以ShuffleNet V2单元为基础,引入ECA（efficient channel attention）注意力模块,使用H-Swish激活函数以便减少网络结构每个Stage模块中ShuffleNet V2单元使用个数,使用轻量化网络设计组件深度可分离卷积。在PlantVillage病害数据集上进行实验。结果表明,模型的参数量约为2.95×105,计算量为3.388×107（FLOPs）和6.674×107（MAdd）,病害识别平均准确率达到了99.24%,为农作物病害识别方法在移动终端等资源受限设备上部署应用提供参考。     

Polymeric pH-sensitive membranes—A review

Significant progress has been achieved in recent years in the field of pH-sensitive membranes. In many cases, pH-sensitive membranes are systems for which the flux, membrane pore size, and solute rejection ratio may be manipulated by changing the pH. This review summarizes recent developments covering the preparation, pH-responsive properties, and applications in various disciplines. The pH-sensitive groups and the evaluation parameters for pH-sensitive membranes are reviewed and discussed. A variety of preparation methodologies, including blending, pore-filling, surface-grafting, and surface-coating techniques are described, and some of their salient features are highlighted. The flat-sheet form and hollow-fiber form pH-sensitive membranes are reviewed. Membrane pore size change and electroviscous effect are discussed. Furthermore, future perspectives of pH-sensitive membranes are discussed.     

管焊缝相控阵超声工程临界评定

介绍相控阵超声检测(PAUT)在工程临界评定(ECA)中的应用。利用已知疲劳裂纹的"退役"试样,通过"亮测"和"盲测",进行PAUT工艺评定,并将PAUT数据与TOFD、MT和VT数据进行比照。结果表明PAUT测评值可信可靠,适于ECA,也符合国际法规ASME BPVC最新版(2013)相关要求。     

Pt/C catalyst degradation in proton exchange membrane fuel cells due to high-frequency potential cycling induced by switching power converters

Proton exchange membrane fuel cells are operated using switching power converters that produce high-frequency ripple currents. These ripples cause high-frequency potential cycling of cells, which is believed to lead premature deterioration in the electrochemical surface area (ECA) of Pt/C catalysts. The qualitative relationship between ECA losses and the frequency of potential cycling was investigated in the range of 1 Hz to 1 kHz. For frequencies higher than 100 Hz, ECA losses were comparable with those at the potential hold condition. However, for lower frequencies, ECA decreased significantly with decreasing frequency. TEM observations showed that there was marked Pt particle growth for the 1-Hz cycling condition, whereas particle size distributions at 100 Hz and potential hold conditions were comparable. The currents associated with Pt oxidation and reduction during potential cycling were also investigated at various potentials and frequencies, and the charges associated with Pt loss (ΔQ) were determined by integrating the measured current. A correlation between the ECA trend and ΔQ was observed. The results obtained in this study are considered informative for electrical engineering research, because it relates to the design of switching power converters that do not negatively influence the Pt/C catalyst durability.     

导电胶的研究进展

导电胶作为无铅连接材料的一种,近年来在电子封装中得到越来越多的重视。导电机理、组成及老化性能的研究成为导电胶实用化的关键因素。各向异性导电胶是连接用Pb/Sn合金的理想替代材料。     

固化工艺参数对导电胶导电性的影响

用四点探针法的原理设计了一种测试导电胶电阻的新方法。研究了固化温度、时间及导电胶层厚度对导电胶固化后导电性的影响。结果表明:夹具经预热导电胶,在80~140℃固化较好;未预热时在110~150℃固化较好,但固化时间较长;随着导电胶层厚度增加,固化时间和固化后的最低电阻都有一定的增大;最佳的固化效果是固化到电阻变化很微小时停止固化。     

Learning algorithm of stage control NBP network

This letter analyzes the reasons why the known Neural Back Promulgation (NBP) network learning algorithm has slower speed and greater sample error. Based on the analysis and experiment, the training group descending Enhanced Combination Algorithm (ECA) is proposed. The analysis of the generalized property and sample error shows that the ECA can heighten thestudy speed and reduce individual error.     

CovaTM及其实现

以原子性、一致性、隔离性和持久性为特征的传统事务模型在满足协同应用的需求方面存在诸多局限。为了解决这一问题,该文提出了一种新的协同事务模型CovaTM,它将协作过程的一次执行看作一个事务,对应的活动看成子事务。CovaTM提供了复杂而灵活的控制流,放松了传统事务的原子性和隔离性的要求,能够支持协作。该文介绍了CovaTM的具体实现机制,包括如何根据协作过程的描述形成事务的各个要素以及运行过程中的按制结构及其算法并与相关工作进行了比较。     

Thermoelectric Behavior of Isotropically Conductive Adhesive

The paper deals with unusual use of one kind of ECA （electrically conductive adhesive）---the ICA （isotropic conductive adhesive）. The main sphere of ECA application is electronic assembly, e.g., it is bonding of semiconductor microchips on printed circuits boards. In this sphere, the ECA compete with soft solder. In spite of this fact, the author utilized of two main ECA characteristics--good electrical conductivity and excellent adhesion to material surfaces to make the fiat thermocouples. Both the design of thermocouples and the measuring device and the measuring workplace arrangement are described. The measured data of thermoelectric voltages are plotted. The thermoelectric （Seebeck＇s） coefficients were calculated from obtained dependences of thermoelectric voltage versus the temperature differences.