小麦生长监测物联网关键技术研究

发布时间：2018-03-31 09:57

本文选题：小麦苗情　切入点：灾害　出处：《中国农业科学院》2015年博士论文

【摘要】：小麦在生长发育过程中有许多内在的规律,其生长过程除了品种因素之外,与气象因子、土壤生态因子等因素高度相关,相应的生长环境条件会产生对应的植株生理和物理特征。因此,探索研究以田间监测数据计算和推理小麦的生长状况成为可能。如何及时、准确和全面的获取小麦生长过程的关键数据,对多源数据进行有效的融合分析,将监测信息由原始的传感器数据提升至决策应用知识,以提高数字化、定量化的监测和诊断能力,是当前研究的难点和核心内容。本文在已初步建设完成的国家小麦监控物联网中心平台的基础上,结合物联网与多源信息融合方法与技术,针对小麦生长环境与长势特征等数据的获取与应用分析环节,进一步深入开展研究。在物联网感知层即数据获取环节,深入研究田间无线传感器组网与网内数据融合关键技术；在物联网应用层即数据应用分析环节,分别深入研究基于点面融合的小麦苗情分类分析方法、基于物联网与模型耦合的小麦气象灾害诊断方法和基于图像识别的小麦叶部病害诊断方法；最后,分别设计开发以上研究的原型系统并集成应用于国家小麦监控物联网中心平台,作为小麦监控物联网的核心组成部分,面向用户提供苗情、灾情、病情应用服务,为小麦生产管理与灾害防控提供决策支持。本文研究的创新性成果主要有以下方面：1)提出了基于田间监测数据的苗情分类动态模拟方法,提出了点面融合方法以实现小麦苗情分类等级的空间分布监测；2)提出了物联网多点、实时监测与传统的小麦气象灾害诊断模型的耦合设计方法,实现了小麦气象灾害的动态诊断与由点到面的区域评估；3)提出了基于图像特征分类识别的小麦病害诊断方法,确立了以颜色、纹理、形状三种特征向量组合和径向基核函数的最优支持向量机分类器,建立了基于现场图像识别的小麦叶部病害远程诊断。最后,基于以上方法与技术分别设计与实现了原型系统,作为小麦监控物联网的核心组成部分集成应用于国家小麦监控物联网中心平台,实现了面向全国小麦主产区用户(包括政府管理部门、农技人员、种植者)提供苗情长势、气象灾害与病害的监测与诊断等综合应用服务。本研究为大规模、跨区域、多测点、多尺度的田间多维信息的及时精准获取,多源信息有效融合与智能化动态诊断分析奠定了研究方法基础,提供了应用案例参考。本文达到了研究预期,具有研究方法的创新意义,实现了良好的实际应用价值。
[Abstract]:There are many inherent rules in the growth and development of wheat. In addition to variety factors, the growth process of wheat is highly related to meteorological factors, soil ecological factors, and so on. Therefore, it is possible to explore the calculation and inference of wheat growth status based on field monitoring data. Accurate and comprehensive acquisition of key data of wheat growth process, effective fusion analysis of multi-source data, upgrading of monitoring information from original sensor data to decision application knowledge, in order to improve digitization, Quantitative monitoring and diagnosis ability is the difficult and core content of current research. This paper combines the methods and techniques of Internet of things and multi-source information fusion on the basis of the preliminary construction of the national platform of the national wheat monitoring Internet of things center. The data acquisition and application analysis of wheat growth environment and growth characteristics are further studied. The key technologies of data fusion in field wireless sensor network and network are studied in the perceptual layer of the Internet of things (IOT). In the application layer of the Internet of things, that is, data application analysis, the classification and analysis methods of wheat seedling situation based on point and surface fusion are studied in depth. The method of wheat meteorological disaster diagnosis based on the coupling of the Internet of things and the model and the method of wheat leaf disease diagnosis based on image recognition. Finally, The prototype system is designed and developed and integrated into the national wheat monitor and control Internet of things center platform. As the core component of wheat monitoring Internet of things, the prototype system is designed to provide the application service of seedling situation, disaster situation and disease condition for users. This paper provides decision support for wheat production management and disaster prevention and control. The main innovative achievements of this paper are as follows: 1) A dynamic simulation method of seedling classification based on field monitoring data is proposed. In this paper, a point-surface fusion method is proposed to realize the spatial distribution monitoring of wheat seedling classification grade. (2) the coupling design method of multi-point, real-time monitoring and traditional wheat meteorological disaster diagnosis model is proposed. In this paper, the dynamic diagnosis of wheat meteorological disasters and the regional assessment from point to surface are realized. A method of wheat disease diagnosis based on image feature classification is proposed, and the color and texture are established. An optimal support vector machine classifier based on the combination of three feature vectors of shape and radial basis function kernel function is used to establish remote diagnosis of wheat leaf diseases based on field image recognition. Finally, a prototype system is designed and implemented based on the above methods and techniques. As the core component of wheat monitoring Internet of things, it is integrated into the platform of national wheat monitoring network of things center. It can provide seedling growth for the users (including government management departments, agricultural technicians, growers) in the main wheat production areas of the country. Comprehensive application services such as monitoring and diagnosis of meteorological disasters and diseases. This study is for the timely and accurate acquisition of multi-dimensional information in the field of large-scale, cross-regional, multi-site, multi-scale, and multi-scale. The effective fusion of multi-source information and intelligent dynamic diagnosis and analysis have laid the foundation of the research method and provided the reference for the application case. This paper has achieved the research expectation, has the innovation significance of the research method, has realized the good practical application value.
【学位授予单位】：中国农业科学院
【学位级别】：博士
【学位授予年份】：2015
【分类号】：S512.1;S126

【参考文献】