Pesticide Residue Digestion Of Several Pesticides On Citrus And Construction Of Prediction Model Of BP Neural Network

In recent years,with the improvement of living standards,consumers’ consumption concepts and dietary attitudes are undergoing tremendous changes.The concept of food consumption has changed from demand to pursuit.This has led to the attention of food safety issues such as pesticide residues in agricultural products.The main reasons for the occurrence of pesticide residues exceeding the national standards in actual life are as follows: excessive application of pesticides,excessive pesticide residues caused by improper application of pesticides;improper drug selection,the residues of selected pesticides are the original pesticides and their toxicity The total residual amount of metabolites and impurities;pesticide characteristics,the characteristics of pesticides themselves cause the residual time to be too long.As an important economic crop of the country,citrus has grown more than apples and ranks first in the country.China’s citrus industry involves everything from nursery to fresh food to processing,with an annual output of more than 40 million tons of citrus and 3.2 million tons of orange juice.Citrus is also a processing raw material for major international agricultural products and export products.According to customs statistics,my country exported 1.03 million tons of citrus products last year,valued at US$1.34 billion.Therefore,the needs of the people and the economic market have urgent needs for research on pesticide residues on citrus.Establishing a predictive analysis system for pesticide residues on citrus can not only provide an effective platform for the public to understand the pesticide residue information on citrus,but also have special guiding significance for citrus farmers to use pesticides.It can also provide a reference for the country to formulate relevant pesticide residue laws.In order to achieve the accurate prediction of pesticide residues on citrus,the following research was carried out.First of all,carry out agricultural residue test.In the field experiment part,five pesticides commonly used in citrus were selected to carry out pesticide residue test in12 regions and nine varieties of citrus.Then indoor detection was carried out in the laboratory,and standard curve was drawn by external standard method.Based on the detection methods of gas chromatography,liquid chromatography and high performance liquid chromatography tandem mass spectrometry,qualitative and quantitative detection and analysis methods of pesticides were established The results showed that the recovery was between 75% and 112%,and the data of the final pesticide residues and the dynamic data of the degradation residues were obtained.Secondly,the BP neural network model is constructed.In the first step,the physical and chemical characteristics of pesticide application,the characteristics of planting environment,and the characteristics of Citrus germplasm were digitized.In the second step,SPSS software was used to analyze the influencing factors.Six common factors were extracted from the citrus model.In the third step,the three-layer neural network model is selected and the parameter operation model is set.After 115 iterations,the citrus model meets the error requirements,the algorithm converges and completes the neural network training.After the final model fitting,the test samples are brought in for prediction.After calculation,the average relative error between the predicted value and the actual value of the citrus test samples is 12.50%.In China,the distribution characteristics of rice are similar to those of citrus,so rice is selected to expand the model.Five common factors were extracted from rice expansion model.In the ninth iteration of rice model training,the algorithm converged to meet the expected error requirements and completed the network training.The average relative error between the actual value and the predicted value in the rice sample is 37.72%,which further shows that the prediction accuracy of the prediction model is high.The BP neural network model on citrus organically combines the pesticide residue data of citrus with the BP network model,so that the pesticide residue data can be modeled and accurately predicted.The model provides a reference value for the prediction of pesticide residues on crops,the rational application of pesticides by farmers,and the country’s formulation of citrus-related pesticide residue regulations.