【作者】 董静；

【导师】 王鹏新； 赵春江；

【作者基本信息】 中国农业大学 ， 农业信息化技术， 2017， 博士

【摘要】 我国是栽培食用菌最早的国家,食用菌产量现己居世界首位,但与国外发达国家相比,仍存在巨大差距,如产业化发展的时间较短,发展不均衡、生产规模偏小、技术水平低、产品质量差、竞争能力弱等问题。随着我国供给侧改革和食用菌产业整体提升的需要,急需一种低成本、专业化服务、推广方便的食用菌信息化解决方案,促进食用菌产业向组织化、规模化、标准化和专业化方向转变,提升我国食用菌产业信息化管理能力。近年来发展的"云计算"技术通过网络提供云端基础设施、平台和软件服务,这种面向服务的新型计算模式为解决食用菌生产和技术推广领域中需求和投入的矛盾,建设低成本、专业化的食用菌业务应用服务系统提供了新思路。本研究主要研究内容和结论如下:(1)设计了食用菌规模化生产云服务平台的架构结合对我国食用菌生产业务特点分析,归纳总结了食用菌规模化云服务平台建构的关键问题;最后针对这些关键问题,设计了食用菌规模化云服务平台的整体架构和核心服务。(2)食用菌物联网多源异构海量数据服务架构研究。在云服务平台中,采用感知数据的符号表示方法,实现了对传感器数据的一致化统一描述和表达,设计了食用菌物联网数据交互协议规范;提出了一种通用性较强的传感器数据注册适配接入机制;引入消息服务总线,采用分布式面向服务构建方法,设计了食用菌物联网数据服务管理平台的整体架构。(3)提出了云环境下基于前向模糊推理的菇房环境预警模型。采用分布图法、加权平均和最优方差等方法,实现了菇房环境监测数据的预处理和融合技术,构建了稳定可靠地感知数据预处理模型;研究基于菇房环境关键因子的模糊化表达,构建了基于多维环境因子的模糊推理模型,针对食用菌云服务平台提供的数据和服务,采用分布式计算模型,优化构建了菇房环境监测预警推理模型。(4)基于矩阵分解离散Hopfield的食用菌干燥控制模型的研究。围绕着基于云服务平台提供的服务,首先通过食用菌介电特性,设计了在线的水分监测框架,构建了食用菌水分测定模型;采用优化链接权重和采用矩阵分解的DHNN模型,构建基于水分在线测定的食用菌干燥控制模型,实现了基于云服务平台的食用菌水分在线监测与干燥控制模型。(5)基于网络新媒体的食用菌规模云服务模式研究。实现智能设备+微信公众号的服务模式,为解决食用菌种植现有推广难、成本高、服务内容少的问题,提供一种新的重要的解决方案。更多还原

【Abstract】 China is the earliest country to cultivate the edible mushroom,now the edible mushroom production is ranked the first in the world,but compared with foreign developed countries,there is still a huge gap,such as the industrialization development time is shorter,uneven development,production scale is too small,low level of technology,poor product quality,weak competitiveness and other issues.With the supply-side reform in China and the need of integral elevation of the edible mushroom industry,it is in urgent need of a low-cost,professional service and convenient promotion edible mushroom information-based integrated solution,to facilitate the edible mushroom industry changed toward to organized,scale,standardization and professional direction and upgrade the information management capabilities of China’s edible mushroom industry.In recent years,the development of "Cloud Computing" technology which provides the cloud infrastructure,platform and software services through the network,this kind of service-oriented new computing model provides a new idea to solve the contradiction between demand and investment in the field of edible mushroom production and technology promotion,and construct the low-cost,professional edible mushroom business application service system.The main contents of this study are as follows:(Ⅰ)Designed the architecture of large scale production cloud service platform of edible mushroom and combined the analysis on the characteristics of edible mushroom production business in China,generalized and summarized the key issues in the construction of edible mushroom scale cloud service platform.Finally,aimed at these key issues,designed the overall structure and core service of edible mushroom scale cloud service platform.(Ⅱ)Edible mushroom IoT multi-source heterogeneous mass data service architecture.In the cloud service platform,using the symbolic representation of perceived data to realize the unified description and expression of the sensor data and design the edible mushroom IoT data exchange protocol specification,proposed a general sensor data registration and adaptive access mechanism.Introduce message service bus,use distributed service-oriented construction method to design the overall structure of the edible mushroom IoT data service management platform,provide the perceptual data and technical service support for various intelligent models and business application systems based on this platform.(Ⅲ)Proposed the mushroom environment early warning model based on the forward fuzzy reasoning under the cloud environment and use the distribution diagram,weighted average,optimal variance and other methods to achieve the pre-processing and integration technology of mushroom house environment monitoring data,and construct the stable and reliable sensor data pretreatment model.The study using the fuzzy expression of the mushroom house environment key factors to construct the fuzzy reasoning model based on multi-dimensional environmental factors.Aimded at the data and services provided by the edible mushroom cloud service platform,use distributed computing model to optimize and construct the mushroom house environment monitoring warning reasoning model.(IV)Study on the edible mushroom dry control model based on matrix decomposition discrete HopField.Around the service provided the cloud-based service platform,first through the dielectric properties of edible mushroom to design an online moisture monitoring framework and establish the edible mushroom moisture determination model.Using the optimized link weight and the matrix decomposition DHNN model to construct the edible mushroom dry control model based on online moisture determination,to realize the edible mushroom online moisture determination and dry control model.(V)Study on the edible mushroom scale cloud-based service model based on new network media.Realize the service model of smart device and WeChat public number,and provide a new and important solution to solve the problems of existing promotion difficult,high cost and less service content of edible mushroom cultivation.更多还原