Research On Water Quality And Soil Quality Based On Neural Networ

For crop planting,once the soil salinization occurs,it will not only affect the yield,but also block the normal development.The salinization of soil means that the whole farming environment becomes worse,and many crops cannot get enough water and nutrients,and their yields are naturally greatly reduced.In addition,salinization can cause the surface of the soil to harden,preventing plant roots from sinking to absorb mineral nutrients.Factors such as the use of different types of irrigation water,climate,rainfall,surface water evaporation,surface and groundwater flow will affect soil salinity.Among them,irrigation water quality has a great impact on the degree of soil salinization.Therefore,it is of practical significance to study soil salinization and irrigation water quality for agricultural production.The main research contents and achievements of this paper are as follows:（1）The pearson correlation coefficient between the various chemical components of the soil is studied,and the relationship between the components is analyzed.Combined with the principal component analysis,the data is reduced in dimension,the calculation amount of the later quantitative analysis is reduced,and the calculation efficiency is improved.Here On the basis,the chemical components in the soil were analyzed from the vertical and horizontal directions;in addition,the relevant chemical indicators of water quality were obtained through experiments,and 18 water sources were analyzed for water quality.（2）Combined with three machine learning algorithms,soil salinity and chemical components of irrigation water were used as independent variables and dependent variables,respectively,and the partial least squares regression（PLS）,support vector machine regression（SVR）and BP neural network were compared and analyzed.（BPNN）The effect of three algorithms in soil salinity prediction.In addition,in order to further improve the fitting accuracy of the model,the genetic algorithm（GA）was used to optimize the four parameters of the BP neural network hidden layer number,hidden layer node number,transfer function and training function,so that the fitted soil contains Predicted salt levels are closer to actual soil salinity.（3）In order to effectively predict the quantitative relationship between soil salinity and chemical components in irrigation water,such as SO₄^2-,K⁺Na⁺,Ca²⁺,HCO₃^-,Mg²⁺,CO₃^2-,salinity,etc.,it is necessary to carry out the main operating conditions in irrigation.optimization.Three types of the most common component factors are used here,namely x,lnx,and^,to establish a prediction model for the corresponding components of soil salinity and irrigation water.Particle swarm algorithm（PSO）was used to optimize the model,and the optimal values of each component of irrigation water quality under the highest accuracy of soil salinity prediction were obtained.Quantitative relationship and rational selection of irrigation water provide reference.