Research On Key Technologies Of Internet Of Things And Development Of Informatization System For Smart Agriculture

Smart agriculture is an important trend and direction of current agricultural development.In the process of agricultural modernization and intelligent development,the Internet of Things technology is used to realize the intelligent perception,intelligent diagnosis and intelligent control of key data,which has important practical value.An important part of the construction of modern agricultural Internet of Things is how to carry out low-cost network transformation of traditional online physical quantity detection devices.At the same time,in the development of smart agriculture represented by agricultural science and technology parks,the current serious leakage in the water pipe network has become another key issue restricting the development of smart agriculture.In view of these status quo,in this dissertation,Io T embedded gateways are researched and developed for smart agriculture,agricultural water pipe network leakage detection systems and methods,and information integrated service supervision platform for smart agriculture parks are researched and developed.Firstly,based on the demand analysis for smart agricultural Io T gateways and pipeline network leakage detection,the overall architecture of an information system for smart agriculture is designed,including general Io T gateways and pipeline network leakage detection systems.Secondly,for the smart agricultural Internet of Things,an embedded gateway software and hardware system based on STM32 has been developed.The system adopts a low-power design method and is powered by solar energy and storage batteries.The system has a multi-mode data acquisition function,which can realize 4-20 m A sensor,pulse sensor and serial communication sensor and other modes of signal acquisition.After the data is processed,the 4G communication module exchanges information with the remote cloud service platform.Thirdly,for the leakage of the pipe network,a diagnosis model is established based on a data-driven method.By analyzing the pseudo-periodical characteristics of water in agriculture,the k-nearest neighbor method is used for data preprocessing.A sliding time window is used to dynamically divide the time period.For the daily corresponding data in the training data,combined with the daytime features,K-means clustering is used to extract the abnormal features of the traffic.The confidence interval corresponding to the confidence level of abnormal flow is set in the system.Combined with the confidence interval,an improved random forest classification method is used to comprehensively judge whether there is an abnormal leakage.Finally,for smart agriculture,a comprehensive information-based service supervision platform was designed and implemented.By interacting with the gateway,the platform realizes data collection,storage and real-time display functions,as well as functions such as park and user management.On the server side,based on the pipeline network leakage detection algorithm,the online pipeline network leakage monitoring and alarm functions are realized.On the client side,data management and chart analysis functions are realized.The system developed in this article has been put into trial operation in March 2021.The system is currently operating stably and reliably,which has significantly improved the information management and service level of smart agriculture in the Agricultural Science and Technology Park,and has good application value.