Cloud Monitoring System For Soldier Fly Breeding Environment Based On Internet Of Things

In recent years,with the continuous improvement of people’s living standards and the rapid development of China’s animal husbandry and catering industry,more livestock manure and kitchen waste have been produced,which has caused huge pollution to the environment.It is found that water flies can feed on livestock manure and kitchen waste,which on the one hand reduces the ecological pressure caused by environmental pollution caused by the accumulation of livestock manure and kitchen waste;On the other hand,as a high-quality protein feed,soldier fly can be used as fish feed to reduce the cost of aquaculture.Therefore,promoting the breeding of water flies can bring rich economic benefits,but the survival rate of water flies and the conversion rate of livestock manure and kitchen waste largely depend on the quality of the environment,so it is necessary to strictly monitor the breeding environment of water flies.The traditional breeding method has the disadvantages of time-consuming and laborious,low efficiency and low real-time performance.In case of emergencies,measures can not be taken in time for real-time regulation.At present,the relevant environmental monitoring systems can not provide remote management function,the form of customer-oriented data visualization is not perfect,and there is a lack of further mining and utilization of environmental data.In view of the above problems,combined with the functional and performance requirements of breeding environment monitoring,this thesis designs a cloud monitoring system of soldier fly breeding environment based on the Internet of things.The system can achieve the functions of obtaining environmental data,visual display,adjusting environmental conditions and early warning.According to the functional requirements of the system,this thesis divides the system into four parts:lower computer,cloud platform,client and prediction algorithm model.Firstly,the lower computer is designed and built.The lower computer completes the collection and processing of environmental data,transmits it to the cloud platform by Ethernet communication,and adjusts the abnormal environment to the best state through relay control environmental equipment.Then complete the design and deployment of the cloud platform background service and client.The cloud platform is responsible for data storage and interaction.It is the central part of the system.The background service and database of the whole remote monitoring system are deployed on the cloud server.Based on B/S mode and Java Web technology,the client realizes the visual display of environmental data,remote management of environmental equipment and early warning function.Finally,in order to deal with emergencies,timely predict the possible environmental data values at the next moment,feed back effective information to the staff for regulation,and prevent the sudden change of environmental parameters from endangering the growth of water flies.This thesis studies the environmental value prediction algorithm,compares the advantages and disadvantages of recurrent neural network(RNN),long short-term memory(LSTM)and gated recurrent unit(GRU),uses the gated cyclic unit network based on one-dimensional convolution to predict the environmental value,and uses cross-correlation fitting to make the prediction result closer to the real value and further improve the prediction ability.After three breeding cycles of water flies,the results show that the system can operate stably,and all functions are normal.It can realize the real-time monitoring and remote management of the breeding environment of water flies,significantly improve the survival rate of water flies and the conversion rate of livestock manure and kitchen waste,and meet the practical requirements.