基于物联网的番茄智能灌溉系统设计

发布时间：2018-06-02 14:05

本文选题：物联网 + 灌溉　；参考：《安徽农业大学》2014年硕士论文

【摘要】：我国水资源人均占有量少、时空分布不均匀、利用困难、水污染加剧,水资源紧迫的现状严重制约着我国的经济的发展。因为农业用水在水资源利用方面所占比例最大,先进的节水灌溉系统的设计可以提高经济效益,改善生态环境,达到节约用水的目的,所以本文从农业灌溉的视角去构思节水途径。本文在基于物联网技术的基础上设计了一套智能灌溉系统。系统首先在感知层ZigBee终端节点采集到需要监测的环境信息,然后通过协调器传送到GPRS上,网络层负责把信息通过GPRS传递到Internet上,并最终传入位于应用层的服务器上,由服务器上设计的处理算法对监测数据进行处理,并根据灌溉策略给出灌溉指令,并由终端节点的执行程序执行。在对整个过程的实际开发中,笔者对各个环节的主要技术进行了深入研究,取得了以下成果：1.文章分别从农田水文循环、作物生理周期等方面研究了作物需水量的影响因素,总结蒸散量的多种计算方法,在综合所有蒸散量影响因子的基础上,利用非线性主成分分析法(NLPCA)和径向基(RBF)神经网络组成的模型(NLPCA-RBF)对番茄蒸散量进行估算。在既保证ET影响因素信息完整,又可消除影响因素之间相关性的前提下,利用NLPCA将影响ET的7个气象因素简化为3个综合成分,并以此为网络训练的输入数据,根据实测的蒸散量作为网络输出建立了RBF神经网络,并且经非训练样本点数据检验。结果表明,与传统RBF网络模型较,NLPCA-RBF网络预测模型能够更好的反应影响因子与蒸散量之间的关系,取得更为精确的结果。2.针对灌溉系统普遍具有的迟滞性和非线性等缺点,本文所设计的灌溉系统引入了模糊控制器作为灌溉策略,相比于单一的输入因素,模糊控制器的输入选择土壤湿度和蒸散量两个因素,这会使得到的灌溉策略更加的精准。最后通过试验验证了模糊控制器的合理性。3.本系统采用模块化的设计思想,以MC9S12XSl28微控制器为核心,加入相应的功能模块完成了ZigBee节点的信息采集模块、通信模块,执行机构的设计,使ZigBee网络能够完成信息的采集、传递,以及灌溉的执行。服务器端采用MVC构架设计,能够强制性的把应用程序的数据展示、数据处理和流程控制分开,提高了开发效率,方便修改。
[Abstract]:The present situation of water resources per capita is less, the distribution of time and space is uneven, the utilization is difficult, the water pollution intensifies, and the urgent water resources seriously restrict the economic development of our country. Because agricultural water use accounts for the largest proportion in the utilization of water resources, the design of advanced water-saving irrigation systems can improve economic efficiency, improve the ecological environment, and achieve the purpose of water conservation. Therefore this article from the agricultural irrigation angle of view to conceive the water saving way. This paper designs an intelligent irrigation system based on Internet of things technology. The system first collects the environment information to be monitored at the ZigBee terminal node in the perception layer, and then transmits the information to the GPRS through the coordinator. The network layer is responsible for transferring the information to the Internet through the GPRS, and finally to the server located in the application layer. The monitoring data are processed by the processing algorithm designed on the server, and the irrigation instructions are given according to the irrigation strategy, and executed by the executing program of the terminal node. In the actual development of the whole process, the author has carried on the thorough research to each link's main technology, has obtained the following achievement: 1. In this paper, the influencing factors of crop water demand are studied from the aspects of farmland hydrological cycle and crop physiological cycle, and the various calculation methods of evapotranspiration are summarized, on the basis of synthesizing all the influencing factors of evapotranspiration. The model of nonlinear principal component analysis (NLPCA) and radial basis function (RBF) neural network was used to estimate the evapotranspiration of tomato. On the premise of ensuring the integrity of the information of the et influencing factors and eliminating the correlation between the influencing factors, the seven meteorological factors affecting et are simplified into three comprehensive components by using NLPCA, which is used as the input data of the network training. According to the measured evapotranspiration as the output of the network, the RBF neural network is established and tested by the untrained sample data. The results show that compared with the traditional RBF neural network model, the prediction model of NLPCA-RBF neural network can better reflect the relationship between the influence factors and evapotranspiration, and obtain more accurate results. 2. In view of the common hysteresis and nonlinearity of irrigation system, the irrigation system designed in this paper introduces fuzzy controller as irrigation strategy, compared with a single input factor. The input of fuzzy controller selects soil moisture and evapotranspiration, which makes the irrigation strategy more accurate. Finally, the rationality of the fuzzy controller is verified by experiments. This system adopts the modular design idea, takes the MC9S12XSl28 microcontroller as the core, and adds the corresponding function module to complete the design of the information collection module, the communication module and the executive mechanism of the ZigBee node, so that the ZigBee network can complete the information collection and transmission. And the execution of irrigation. The server is designed with MVC architecture, which can force the application data display, data processing and flow control, improve the development efficiency and easy to modify.
【学位授予单位】：安徽农业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：S274;TP391.44;TN929.5

【相似文献】