基于键合图理论的间歇式精准喷施系统喷雾压力动态特性

为解决传统喷施系统农药利用率低的问题,设计基于双阀控制的精准间歇式喷施系统。针对喷施系统多能域相互耦合的特点,采用键合图方法建立系统状态方程,并基于20-sim键合图仿真平台进行仿真计算,得到不同压力工况下的喷雾压力动态特性。试验研究喷施系统在0.2、0.3和0.4MPa压力下喷头的喷雾压力动态特性,并采用轮廓分析统计学方法研究仿真结果的准确率。试验结果表明:车速为0.8m/s,作物间距20cm,作物茎叶覆盖直径10cm情况下,间歇式精准喷施系统的喷雾准确率89%,仿真计算与试验结果最大相合置信水平为0.088。基于键合图理论建立的数值模型能够很好的预测间歇式精准喷施系统的喷雾压力动态特性。     

Identiifcation of microRNAs in two species of tomato,Solanum lycopersicum and Solanum habrochaites,by deep sequencing

MicroRNAs (miRNAs) are ~21 nucleotide (nt), endogenous RNAs that regulate gene expression in plants. Increasing evidence suggests that miRNAs play an important role in species-speciifc development in p...     

An examination of the salmon aquaculture conflict in British Columbia using the graph model for conflict resolution

In 1995, the British Columbia Government imposed a moratorium on salmon farming expansion and conducted an environmental assessment review of the industry. Despite the generally favourable conclusions of this review, the government delayed making a decision on the future of the industry for more than 2 years. This issue is examined using a graph model for conflict resolution. The graph model requires only ordinal preferences for the decision‐makers and allows the analyst to impose realistic constraints on the conflict such as providing for unidirectional and common moves. Given the preferences of the players involved in the conflict, inaction on the part of the provincial government was a stable solution for 2 years. The deteriorating economic situation in British Columbia combined with an improved public acceptance of the salmon farming industry shifted the solution to one of adopting the recommendations from the environmental assessment review and eventually lifting the moratorium.     

秦代工程技术语言初探

考查和测绘了秦陵铜车马及在兵马俑坑出土的构件，在史料分析及实物考证的基础上，研究了秦代工程技术语言产生的条件，探讨了以文字式、平面几何图形、立体图形、实物或模型式以及多面图形等五种秦代工程技术语言的表达形式．     

无人机多光谱遥感反演各生育期玉米根域土壤含水率

为准确及时地获取植被覆盖条件下农田土壤水分信息,该文以不同水分处理的大田玉米为研究对象,利用无人机遥感平台对夏玉米进行多期遥感监测,并同步采集玉米根域不同深度土壤含水率(Soil Water Content,SWC)。基于2018年夏玉米拔节期、抽雄-吐丝期和乳熟-成熟期的无人机多光谱遥感影像数据集,通过支持向量机(Support Vector Machine,SVM)分类剔除土壤背景,提取玉米冠层光谱反射率并计算10种植被指数(VegetationIndex,VI),然后利用全子集筛选(FullSubsetSelection)法对不同波段和植被指数进行不同深度土壤含水率的敏感性分析,并分别采用岭回归(Ridge Regression,RR)和极限学习机(ExtremeLearningMachine,ELM)2种方法构建全子集筛选后0～20、20～45和45～60cm不同深度下的土壤含水率定量估算模型。结果表明:基于贝叶斯信息准则(BayesianInformationCriterion,BIC)的全子集筛选法可以有效筛选最优光谱子集,筛选变量基本都通过了显著性检验,自变量个数较少;在同一生育期、同一深度条件下,ELM模型效果均优于RR模型;玉米在拔节期、抽雄-吐丝期的最佳监测深度为0～20cm,在乳熟-成熟期的最佳监测深度为20～45cm;乳熟-成熟期的20～45cm深度下的ELM反演模型效果最优,其建模集和验证集的决定系数Rc2和Rv2分别为0.825和0.750,均方根误差RMSEc和RMSEv分别为1.00%和1.32%,标准均方根误差NRMSEc和NRMSEv分别为10.85%和13.55%。利用全子集筛选法与机器学习相结合的方法可以提高土壤含水率的反演精度和鲁棒性,本研究为快速、准确地监测农田土壤墒情、实施精准灌溉提供了一种新的途径。     

秸秆深层覆盖对水分入渗及蒸发的影响

 为探讨秸秆还田对土壤水分入渗和蒸发造成的影响,通过在土表下20 cm处铺设3cm厚秸秆隔层的土柱试验表明:在灌溉水分入渗时,秸秆隔层阻碍了土壤重力水入渗,降低了土壤重力水由土壤上层向下层入渗的速率;在土壤水分蒸发过程中,秸秆层隔断了土壤的毛细管,使秸秆层以下土壤水很难通过土壤毛细管作用而向地表运移并蒸发。秸秆深层覆盖可以使深层土壤水分的蒸散量减少2%～3%,对深层土壤的蓄水保墒有积极的作用。     

基于深度学习的水产病害可视化知识图谱构建与验证

知识图谱本质上是基于图的语义网络，表示实体与实体之间的关系，在知识问答、语义检索等领域起着至关重要的作用。针对目前水产病害领域存在实体关系交叉关联、多源异构数据聚合能力差、利用率低、知识共享困难等问题，该研究基于自然语言处理和文本挖掘提出了一个基于神经网络深度学习模型的水产病害专业领域知识图谱构建方法并进行试验验证。首先，构建水产病害专业领域本体，并预定义实体类型、属性和关系的集合，确定知识抽取边界；其次，在本体基础上，分别利用规则方法和深度学习方法对半结构化和非结构化知识进行抽取。对于非结构化知识，提出“水产病害+关系+BMES”文本标注体系，将关系抽取融合于命名实体识别任务中直接对三元组建模，将实体关系抽取转化为序列标注问题，不仅提高标注效率，还实现了实体和关系的联合抽取。同时通过标签匹配和映射对三元组建模获得RDF数据，解决了重叠关系抽取的难题。利用BERT-BiLSTM+CRF端到端模型进行试验，试验结果证明该三元组抽取方法具有较高的召回率（89.64%），准确率（94.04%）和F₁值（91.34%），优于CNN+BiLSTM+CRF和BiLSTM+CRF等模型，抽取效果有了显著提升，并将抽取到的知识存储到 Neo4j 图数据库中，实现知识可视化管理及知识推理分析。该研究构建的水产病害知识图谱精度高、粒度细，能够帮助机器理解数据、解释现象、知识推理，从而发掘深层关系、实现智慧搜索与智能交互。     

棉花和瓜类上棉蚜相互转换后取食行为的EPG研究

实验采用 EPG技术测定了棉花和 3种瓜类上棉蚜转换后的取食行为 ,结果表明 ,西瓜、黄瓜和南瓜上的棉蚜转接到棉花叶片上取食 ,均可出现 A、B、C、E、F和 G6种取食波 ,其中寻食刺探的次数和时间均较少 ,且每次达到韧皮部吸取营养的时间长于 1 0 min。而棉花上的棉蚜转到黄瓜叶片后 ,寻食和刺探的时间占整个处理时间的95 %以上 ,口针到达韧皮部的次数少、持续时间很短。黄瓜叶片不利用于棉花上棉蚜的取食     

论高职高专《单片机原理及应用》课程改革

分析了高职《高专单片机原理及应用》课程教学存在的问题,根据实践教学经验进行了课程改革探讨,提出了通过改变授课顺序、降低课程难度、知识结构逐步深入、培养学生创新能力、采用多种教学方法,从多方面入手,增强学生的学习兴趣,全方位提高学生的实践动手能力。     

Dynamics of nitrogen and phosphorus in paddy field under deep application condition of urea pill and its pollution-reducing potential

Dynamics of the pollutants such as nitrogen, phosphorus and the suspended particulates in the field water and soil and its pollution-reducing potential are investigated in the rice plant returning-green stage using deep application technique of urea pill with the application depth of 3, 7 and 11 cm (expressed as t-3, t-7 and t-11, respectively) based on the micro-zone simulation field experiments. Results show that nitrogen content in the field water gradually decreases within 5~7 d after deep allocation of urea, and reaches the low level after the seventh day. NO-3-N concentration is higher than NH+4-N, but there is little difference between treatments. As the pollution source is concerned, NO-3-N in the field water should be considered. There isn’t any correlation between nitrogen content in the water and urea application depth in the field. By contrast with blank treatment（expressed as t-c）, NO-3-N, NH+4-N and TN contents in the field water with deep application of urea pill decrease 29%～47%, 64%~89% and 79%～97%, respectively. The TP and DP are in the high level in the field water with t-3 and t-7. The TP, DP and SS concentration in the field water with t-11 decrease 32%～73%, 92%～99% and 50%～80%, suggesting that deep application of urea pill into 11 cm layer of the field should have a significant pollution-reducing potential and become one of effective measures to control non-point source pollution of agricultural system.