Research Of Intelligent Aquaculture System Based On Big Data Technology

With the development and popularization of Internet of things,aquaculture data storage is expanding rapidly,which is an opportunity and challenge for aquaculture industry.The existing aquaculture information platform provides a relatively complete data display and retrieval service,but the data are independent of the relationship and value of the data,and the value is not maximized.In order to give full play to the value of aquaculture data resources,and to solve the problem that aquaculture information is low,data processing efficiency is low,and the correlation relationship of aquaculture factors is difficult to obtain,this paper designs a intelligent aquaculture system based on Hadoop distributed architecture.The large data of aquaculture is stored in the system,and the data are analyzed by the scheme,and the multi factor relation model is generated and the trend of data change is predicted.Finally,the information service related to aquaculture is provided to the culture and scientific researchers.In this paper,based on the popular Hadoop data frame,the four layer architecture of aquaculture large data platform is designed according to the production process and industry characteristics of aquaculture.In view of the complex aquaculture environment and the difficulty of obtaining the relationship between various water elements,the utilization of the aquaculture factors is strong.The feedforward error back propagation learning(BP neural network)algorithm is used as a data analysis method to solve the analysis of complex nonlinear relationship between aquaculture elements.On the basis of this,the analysis model of aquaculture elements is proposed and the Johnson attribute reduction algorithm based on discernable matrix matrix is introduced to the traditional calculation.The method is optimized and the convergence speed of the network is improved when the precision is guaranteed,and then the MapReduce distributed programming model is used to parallelize the BP neural network algorithm to realize the mass data processing demand of the aquaculture large data platform,and the analysis model of aquaculture elements and the parallel learning algorithm are analyzed.According to the annual output prediction of Shandong prawn and the classification evaluation of aquaculture water quality as an example,the detailed design and result analysis were carried out.Finally,the large data structure design and data analysis method were integrated to build a large data system of high efficiency and high error tolerant aquaculture data management,mining and visualization of aquaculture.In this paper,an excellent software development technology is used in the research and development.In view of the demand of aquaculture production,the intelligent aquaculture system,which integrates data storage,data management and analysis and evaluation of aquaculture elements,is designed and realized.The system module can provide aquaculture information management,multi factor prediction analysis and efficient mass evaluation of water quality.It has a certain value for data management and data analysis and mining of mass aquaculture.