Research on Key technologies of soil moisture monitoring and intelligent irrigation cloud service platform construction

【作者】 李淑华；

【导师】 周清波；

【作者基本信息】 中国农业科学院 ， 农业遥感， 2016， 博士

【摘要】 “云计算”主要通过互联网为用户提供基础设施、平台和软件服务,这种面向服务新型计算模式为解决当前农田土壤墒情监测和灌溉决策面临的信息服务需求与投入之间的矛盾问题提供了新的思路。通过云服务平台,可以有效降低用户使用墒情监测信息指导灌溉决策的服务成本和系统建设投入风险,提高节水灌溉管理信息化建设效率。通过按需服务,快速集成构建应用系统是实现云服务中“软件即服务(Soft as a service,Saas)”的关键,针对这一关键问题,本研究以快速高效构建农田土壤墒情监测与智能灌溉系统为目标,研究提出适合土壤墒情监测与智能灌溉决策服务的云服务平台体系架构,发展农田土壤墒情监测云服务粒度优化设计方法,提出基于服务的农田土壤墒情监测与智能灌溉应用系统快速构建技术方法,并对墒情监测和智能灌溉决策中的关键云服务进行设计和实现,为构建高效安全的农田土壤墒情监测与智能灌溉云服务平台提供理论和技术基础。本研究主要研究内容及结论如下:(1)分析了云服务平台体系结构及与之相适应的的关键技术问题。通过对比分析主流云服务平台的体系结构特点,结合农田土壤墒情监测与智能灌溉业务领域分析,确定了私有云是目前搭建墒情监测与智能灌溉云服务平台的最有效方式,且服务粒度设计、服务管理与集成以及安全性问题是私有云服务平台构建的技术关键。结合云服务平台框架设计、需求分析和功能设计,设计了适合农田土壤墒情监测与智能灌溉私有云服务平台构建的云服务平台体系结构,这种面向服务的结构在扩展性和用户透明性方面有提升。(2)研究了农田土壤墒情监测与智能灌溉决策云服务粒度优化设计方法。服务粒度作为表征云服务质量的重要属性,很大程度上影响按需服务的灵活性和服务集成的难度。本研究通过综合分析农田土壤墒情监测服务在各特征空间上的相似程度与关联、依赖关系,基于加权有向图,构建了面向整个农田土壤墒情监测业务领域的原子服务相关距离模型;同时,把粒度优化问题转换为路径优化问题,并结合蚁群优化聚类算法提出了云服务粒度优化方法,解决了云服务平台的服务粒度的定量化优化设计问题。通过设计农田土壤墒情监测服务粒度优化实验,引入基于Qos的服务评价模型对比分析了服务粒度优化结果,结果表明通过粒度优化后的服务集合的综合服务粒度适中,更符合高聚合、低耦合和低复杂度的服务粒度设计原则,相比细粒度集合和粗粒度集合,粒度最优后的云服务能提供最高的服务质量。(3)研究了基于服务的农田土壤墒情监测与智能灌溉应用系统快速构建技术方法。服务集成技术水平决定了基于按需服务构建农田土壤墒情监测与智能灌溉决策应用系统的速度和效率。通过分析服务的组织管理模式,并构建服务管理平台提高基础服务的组织运行效率;在分析墒情监测与灌溉控制应用系统常用设备和功能的基础上,基于Web前端设计和开发了用于应用系统定制开发的可视化系统框架,便于应用系统的快速高效构建实现。(4)数据传输服务、灌溉决策分析服务以及安全控制服务研究。安全性、可用性以及服务性能是体现云服务平台质量水平的重要方面。本研究针对数据通讯实时性问题,设计并开发实现了基于观察者模式的实时数据推送机制,可以较低成本实现实时性能的提升;设计了精量灌溉决策云模型库,利用云模型库的应用接口提供优选建模功能,可有效降低因由模型多样性而带来的建模复杂程度,解决了灌溉决策模型的优选与适配问题;针对云服务应用环境中安全访问控制问题,研究提出了一种新的基于角色和空间属性的细粒度的访问控制模型REBAC,该模型具有更强的客观世界表达能力与灵活性,实现对服务细粒度的、多层级及动态限定访问空间区域范围的的访问控制。(5)在以上关键技术研究的基础上,设计农田土壤墒情监测与智能灌溉云服务应用原型系统,并实现其核心功能模块,给出其在全国土壤墒情估计中的应用实例。结果表明,按需云服务方式构建应用系统效率高,系统灵活性和可扩展性强。

【Abstract】 "Cloud computing" as a service-oriented computing model provides infrastructure, platforms and software service for users mainly through the Internet, which offers a new method to solve the contradictions between the demand and investment of information service in the monitoring of soil moisture content and irrigation decision-making of the current farmland. Through the platform of cloud service, the service cost when customers use the information of moisture monitoring to conduct irrigation decision and the investment risk of system construction can be reduced effectively; Moreover, the informationization construction efficiency of water-saving irrigation management can be improved. The rapid integration and construction of the application system through on-demand service, is the key to realize the "software as a service” in cloud service. In order to realize the goal of the rapid construction of farmland soil moisture monitoring and intelligent irrigation system, this study proposes the architecture of cloud service platform suitable for soil moisture monitoring and intelligent irrigation decision-making service, develops an optimization design method of cloud service granularity for farmland soil moisture monitoring, suggests the rapid construction technological methods of farmland soil moisture monitoring and intelligent irrigation application system based on the service, designs and implementation the key cloud service of soil moisture monitoring and intelligent changing decision-making, and provides theoretical and technical basis for the high efficient and safe construction of the farmland soil moisture monitoring and intelligent irrigation cloud service platform. The main conclusions of this study as follows:(1) Analyzed the architecture of cloud service platform and the related key technical issues. Through comparing architecture characteristics of mainstream cloud services platform and combining with the analysis of farmland soil moisture monitoring and intelligent irrigation business field, the private cloud was determined as the most effective way to build the cloud services platform of the moisture content monitoring and intelligent irrigation, and services granularity design, service management and integration, and security issues was the key technology of platform construction for private cloud service. Integrated with the cloud services platform framework design, requirement analysis and function design, architecture of cloud services platform constructed by private cloud services platform suitable for farmland soil moisture monitoring and intelligent irrigation was designed, which had good scalability and user transparency.(2) Researched on the optimization design method of cloud service granularity for farmland soil moisture monitoring and intelligent irrigataion. As an important attribute to characterize the quality of cloud services, the service granularity greatly affected the flexibility of on-demand service and the difficulty of service integration. Through the comprehensive analysis of the similar degree and correlation, dependency of farmland soil moisture monitoring service in the feature space, the atomic services related distance model was constructed to orient the business of farmland soil moisture monitoring services based on weighted directed graph. While design optimization approach for quantitative of service granularity in cloud service platform was solved by transforming granularity optimization problem into a path optimization problem, and combines with the ant colony optimization clustering algorithm to proposed a cloud services granularity optimization method. The result of services granularity optimization experiment of farmland soil moisture monitoring based on the comparative analysis of the QoS service evaluation model showed that comprehensive service granularity of the service granularity optimization set had moderate size, and more in line with high polymerization, low coupling and low complex degree of service granularity design principles. Compared with the set of fine-granularity and coarse-granularity, the cloud service with optimal granularity could provide the highest quality of service.(3) Researched on the rapid construction techniques for farmland soil moisture monitoring and intelligent irrigation system based on service. Service integration technology level determines the construction speed and efficiency of farmland soil moisture monitoring and intelligent decision-making of irrigation application system based on on-demand service. The organization operational efficiency of basic services was improved through analyzing the organization management mode of services, and constructing the service management platform. Analyzed common equipment and the function of soil moisture monitoring and irrigation control application system, the framework of visualization system used for the customization and development of applications system was designed and developed based on web front-end, which was helpful to realize the rapid and efficient construction of application system.(4) Researched on data transmission services, analysis services for irrigation decision-making and security control services. Security, availability and service performance are important aspects of the quality level of cloud service platform. In this study, real-time data push mechanism based on observer mode was designed and developed, aiming at the problem of real-time data communication, which could achieve the improvement of real-time performance in a low cost. Cloud model library of precision irrigation decisionmaking was designed, which could provide optimization-modeling capabilities though using the application interface of cloud model library, effectively reduce the modeling complexity caused by modeling diversity, and solve the optimization and adaptation of irrigation decision-making model. Aiming at the problem of security access control in the cloud service application environment, a new access control model REBAC based on role and spatial attributes fine granularity was proposed, which had stronger expression ability on the objective world and flexibility, and realize the access control on services fine granularity, multi-level, and dynamic limited access to the space area.(5) Based on these key techniques, farmland soil moisture monitoring and intelligent irrigation application of cloud services prototype system was designed, and the application of the core function module was achieved, and the application case in the soil moisture estimation of countrywide was also given in the study. The results showed that the application system constructed on on-demand service had high efficiency, flexibility and extensibility.

【关键词】 云服务； 粒度优化； 数据推送； 墒情监测； 智能灌溉；
【Key words】 Cloud service； Granularity optimization； Data push； Soil moisture monitoring； Intelligent Irrigation；