防治稻瘟病的新型内吸性杀菌剂氰菌胺

概述了防治稻瘟病的新型内吸性杀菌剂氰菌胺，内容涉及创制经纬、化学名称、理化性质、毒性、作用机理、专利与登记、合成方法与应用等。     

基于流形学习的人体动作识别

目的 提出了一个基于流形学习的动作识别框架,用来识别深度图像序列中的人体行为。方法 从Kinect设备获得的深度信息中评估出人体的关节点信息,并用相对关节点位置差作为人体特征表达。在训练阶段,利用LE（Lalpacian eigenmaps）流形学习对高维空间下的训练集进行降维,得到低维隐空间下的运动模型。在识别阶段,用最近邻差值方法将测试序列映射到低维流形空间中去,然后进行匹配计算。在匹配过程中,通过使用改进的Hausdorff距离对低维空间下测试序列和训练运动集的吻合度和相似度进行度量。结果 用Kinect设备捕获的数据进行了实验,取得了良好的效果;同时也在MSR Action3D数据库上进行了测试,结果表明在训练样本较多情况下,本文方法识别效果优于以往方法。结论 实验结果表明本文方法适用于基于深度图像序列的人体动作识别。     

黏膜佐剂mLT63及CpG-ODN鼻内免疫的佐剂作用

目的研究大肠杆菌不耐热肠毒素无毒突变体mLT63和CpG-ODN鼻内免疫对抗原的佐剂效应。方法选用破伤风类毒素(TT)和小牛血清白蛋白(BSA)为抗原,mLT63和CpG-ODN为佐剂,经鼻内免疫BALB/c小鼠。ELISA间接法检测免疫小鼠血清和肺、直肠及阴道组织萃取液的特异性抗TT、BSA的IgA、IgG水平。结果mLT63和CpG-ODN均能协助TT和BSA抗原在血清、肺、阴道、直肠组织中诱导出高滴度的IgG抗体。在血清中IgA抗体滴度均比较低,而TT-IgA在肺和阴道组织中滴度较高,BSA-IgA在肺组织中滴度较高。各佐剂组之间IgG和IgA抗体水平相比,差异均无显著意义。不同剂量的mLT63诱导的BSA-IgA抗体水平差异无显著意义。结论mLT63和CpG-ODN经鼻内免疫对TT和BSA抗原均有良好的佐剂作用。     

基于高精度流线生成的交互流场可视化

高效准确的流线绘制一直是流场可视化的重要研究内容,流线可以对流场的重要特征进行有效的稀疏表示,但流线需要长期的粒子追踪过程及大量的积分计算,在面向大规模流场可视化时时间效率较低,需要高性能计算设备进行辅助计算.本文通过设计一种基于深度学习的高精度流线生成算法,将初始的低精度流线快速映射为稠密的高精度流线,可以在较短的时间内快速生成可靠的流线可视化结果,并在此基础上设计了交互式实时流场可视化系统,涵盖了流场的特征检测,属性关联分析,信息论分析等,帮助用户快速了解流场数据,找到自己感兴趣的区域进行后续进一步深度分析,避免了获取过多冗余数据,同时优化了分析工作的效率,满足用户对于流场结构,特征属性等多维度进行关联分析的需求.     

用于单张图像去雨滴的轻量级网络

图像去雨是图像低等级任务中的热点问题,去雨滴又是图像去雨中很重要的一种情况,附着在玻璃或相机镜头上的雨滴会显著降低场景的可见性.因此,去除雨滴将有助于许多计算机视觉应用,特别是户外监控系统和智能驾驶系统.本文提出了一种用于单张图像去雨滴的轻量级网络算法(PRSEDNet),该网络算法采用递归计算,运用卷积长短期记忆网络(Convolutional LSTM network)和特征提取模块来提取特征,通过与原图像结合来去除雨滴,最终获得高质量的无雨滴清晰图.实验结果表明,我们的PRSEDNet与现有的基于深度学习的去雨滴算法相比,在能达到高效的去雨滴性能的同时,有更少的参数量且计算效率高.     

基于多季相Sentinel-2影像的白洋淀湿地信息提取

白洋淀湿地是华北平原上重要的浅水湖泊湿地,对雄安新区绿色发展具有重要的生态价值。对白洋淀高度异质化的景观格局进行分类,能够为白洋淀湿地资源的遥感监测提供指导意义。针对湿地季节变化的特点,对白洋淀每个季节选取一期具有代表性的Sentinel-2影像,采用分类与回归树（CART）、支持向量机（SVM）、随机森林（RF）3种常用的机器学习分类器对15种季相组合实验方案进行分类,分析不同季相遥感影像及其组合对白洋淀湿地信息提取的优劣。结果表明：相较于使用单一季相影像分类,多季相影像的组合能够显著提高分类精度,春&夏季相组合能够得到最优的分类效果,相对单季影像总体分类精度提高了10.9%~25.5%,Kappa系数提高了0.09~0.29;SVM分类器的分类表现较为稳定,能够得到最高的平均分类精度,CART分类器在处理高维特征的能力不如随机森林和SVM;不同特征类型对湿地信息提取的贡献度从高到底依次是红边光谱特征、传统光谱特征、缨帽变换特征、主成分分析特征、纹理特征。实验成果能为湿地信息的遥感识别提供依据。     

Characteristics of cationic Red X-GRL adsorption by diatomite tailings

As secondary mineral resources, diatomite tailings(DT) got from the Linjiang region of China were prepared and characterized by SEM, XRF and XRD. Mono-factor experiments were carried out to investigate the effects of the operation factor, including contact time, adsorbent concentration, initial p H value of the dye solutions, adsorption temperature and initial concentration of cationic Red X-GRL(X-GRL) on the adsorption of X-GRL. The adsorption kinetics, isotherms, thermodynamics and mechanisms for X-GRL were also studied. It was efficient for DT to adsorb X-GRL from aqueous solutions, and it was even discovered to have higher adsorptivity for X-GRL than diatomite concentrate(DC) in our previous test. The adsorption processes fit very well with the pseudo-second-order model and the Langmuir isotherm equation. In addition, various thermodynamic parameters, such as standard Gibbs free energy(ΔG°), standard enthalpy(ΔH°) and standard entropy(ΔS°) have been calculated. From thermodynamic studies, it was seen that the adsorption was spontaneous and endothermic. The main driving forces of the physical adsorption on DT are electrostatic attraction. The reason why DT showed higher adsorptivity for X-GRL than DC was that there were more clay mineral particles within, which has a remarkable ability of dye adsorption due to its high surface area. DT as a cheap absorbent for X-GRL removal would replace or partially replace the activated carbon.     

Effects of LiAl-layered double hydroxides on early hydration of calcium sulphoaluminate cement paste

As a 3D micro-nano material, layered double hydroxides have been widely used in many fields, especially for reinforced composite materials. In this paper, LiAl-LDHs was obtained by a hydrothermal method. In order to investigate the effects of LiAl-LDHs on the early hydration of calcium sulphoaluminate (CSA) cement paste, compressive strength, setting time and hydration heat were tested while X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FT-IR), scaning electron microscopy (SEM) and differential scanning calorimetry (DSC) analysis were employed. The results indicated that LiAl-LDHs could significantly improve the early compressive strength and shorten the setting time of calcium sulphoaluminate cement paste with 3wt% concentration. Besides, the hydration exothermic rate within 5 h was accelerated with increasing LiAl-LDHs content. Moreover, the addition of LiAl-LDHs did not result in the formation of a new phase, but increased the quantity of hydration products providing higher compressive strength, shorter setting time and denser microstructure.     

Multi-objective robust design optimization of a novel negative Poisson’s ratio bumper system

Negative Poisson’s ratio (NPR) structure has outstanding performances in lightweight and energy absorption, and it can be widely applied in automotive industries. By combining the front anti-collision beam, crash box and NPR structure, a novel NPR bumper system for improving the crashworthiness is first proposed in the work. The performances of the NPR bumper system are detailed studied by comparing to traditional bumper system and aluminum foam filled bumper system. To achieve the rapid design while considering perturbation induced by parameter uncertainties, a multi-objective robust design optimization method of the NPR bumper system is also proposed. The parametric model of the bumper system is constructed by combining the full parametric model of the traditional bumper system and the parametric model of the NPR structure. Optimal Latin hypercube sampling technique and dual response surface method are combined to construct the surrogate models. The multi-objective robust optimization results of the NPR bumper system are then obtained by applying the multi-objective particle swarm optimization algorithm and six sigma criteria. The results yielded from the optimizations indicate that the energy absorption capacity is improved significantly by the NPR bumper system and its performances are further optimized efficiently by the multi-objective robust design optimization method.     

一种水下群机器人路径规划算法的形式化研究

为了增强基于遗传算法的水下群机器人路径规划算法正确性的说服力,使用定理证明对其进行形式化研究,给出算法在定理证明器HOL4中的形式化模型。基于算法形式化的一般步骤,首先对算法的设计进行了详细的分析,指出算法设计的核心步骤与建模难点。在此基础上建立了总体形式化建模框架,然后对其进行化简,得到种群初始化、选择、交叉三个核心模块。接着给出模型中要用到的基本数据类型的形式化描述,并分别对三个模块进行形式化描述,最终得到算法的形式化模型。通过证明与模型相关的97条性质,说明了模型的合理性及有效性,在此模型的基础上,可以完成对算法的形式化验证,同时还能拓展HOL4的应用范围。