Research And Design Of Red Tide Auxiliary Analysis System

In recent years,the occurrence of red tide disasters in China’s coastal areas has become more and more frequent,which has brought huge economic losses to the fishery and Marine aquaculture industry,and seriously destroyed the balance of Marine ecosystem.Therefore,it is very important to find out the red tide in time and give corresponding control measures to reduce the harm brought by the red tide.In this thesis,a red tide auxiliary analysis system is designed with a variety of machine learning algorithms as the core,which provides auxiliary decision-making information for red tide identification and early warning.The specific content is as follows:Firstly,the concept and harm of red tide are introduced,and the significance of red tide warning is clarified.The research status of monitoring and forecasting methods of red tide is also analyzed.Secondly,the physical and chemical factors and biological factors related to the occurrence of red tide are analyzed and summarized,on this basis,two kinds of judgment basis for red tide analysis are given,that is,red tide is judged by analyzing the physical and chemical factors of water quality and red tide is judged by analyzing whether the biomass of red tide reaches the threshold value.Thirdly,according to the red tide judgment basis based on the physical and chemical factors of water quality,the problem of ocean red tide analysis is converted into a classification problem,and decision tree and support vector machine are selected to carry out red tide analysis and modeling on the same sample data set.On this basis,the support vector machine is optimized and the red tide analysis and modeling are carried out by using genetic algorithm.The experimental results show that the accuracy of the three models can reach more than 83%,which proves the feasibility of the three red tide classification models based on the physical and chemical factors of water quality to analyze and judge red tide.Then,according to the red tide biomass based on the red tide judgment basis,the RBF neural network and the generalized regression neural network are selected to establish the relationship model between the red tide biomass and its environmental factors for red tide analysis,on this basis,the particle swarm optimization of the generalized regression neural network and the red tide analysis modeling.The experimental results show that the root mean square error of the three models is less than 24,which proves that the three models are feasible to predict the biological density of red tide.Finally,respectively,the auxiliary analysis system of algal bloom of input,processing,output,should have the function of the three aspects carries on the analysis,based on the designed for red tide auxiliary analysis system user module,the data preprocessing module,model training module and red tides aided analysis module four main modules,and the process of realization of each main function module is given.