Design And Application Of Cloud Platform For Prediction Of Breeding Pig Growth Performance Based On Deep Learning

In the development process of animal husbandry,breeding research has always been the most important link to improve production efficiency.In the previous breeding research process,collecting and analyzing breeding data is a long-term and time-consuming work.Nowadays,with the rapid development of communication infrastructure and the popularization of information technology,more and more enterprises and research institutions engaged in breeding have accepted the application of information technology in breeding work.Therefore,it is very important for the breeding industry to design an IOT cloud platform that can help breeding institutions improve the collection efficiency and use deep learning technology to predict the growth performance of breeding pigs.The design in this thesis implements a cloud platform for predicting the growth performance of breeding pigs.The platform collects the measurement data of breeding pigs through the Internet of Things,and on the basis of the data,deep learning technology is used to realize the function of predicting the growth performance of breeding pigs.The platform is designed as a front-end separation architecture and deployed in a clustered manner.JAVA is used in the back-end,Spring Boot under the JAVA ecosystem is used as the development framework to complete the development of back-end communication and business functions.Vue.js ecological technology will be used in the front-end as the front-end development framework.MQTT will be used in IOT communication end to realize the communication between the device and the back-end.In the data storage phase,Redis+Mysql is used to achieve caching and data persistence.In the data analysis stage,the box chart is used to preprocess the data,and then GRU,LSTM and linear regression models are used to train the processed data.Finally,the pig growth prediction model is obtained.This thesis mainly completes the following contents:(1)Build the IOT data collection platform,design the IOT cloud mode of communication architecture,and achieve automatic collection of breeding pig growth data by using gateway+MQTT.(2)Build distributed Saas service application by using cloud computing technology to provide centralized data storage,analysis and application for pig growth measurement data.(3)Test and analyze the measured growth data by using GRU,LSTM neural network and linear regression model.Analyze the characteristics of the collected data that have an impact on the growth performance of breeding pigs,select the characteristics with larger impact factors for data preprocessing,and then add them to the model for training.The experimental results show that the three models have high accuracy in predicting the growth performance of breeding pigs,and the prediction accuracy of GRU and LSTM is close to each other,and both are higher than the prediction accuracy of linear regression.