Research On Field Irrigation Method Based On Soil Moisture Prediction

The distribution of fresh water resources in our country is uneven,and the utilization rate of agricultural water is low.Jilin Province,as a major food planting province,has a large proportion of farmland in areas where water resources are relatively scarce.How to use water rationally and irrigate is one of the solutions to the high and stable grain production in our province.Big problem.In this paper,aiming at the optimization strategy of irrigation management for the main grain crop corn planting process in our province,this paper proposes a field crop irrigation method based on machine learning.This method uses machine learning algorithms to predict future soil moisture change trends,and the results are used as the basis for field irrigation strategies,and finally water-saving irrigation is realized.Considering that the current short-term prediction of soil moisture mainly uses meteorological factors and time series,this paper proposes a prediction model based on the BAS-BP neural network to solve the problems of high input dimensions,insufficient accuracy,and defects in the BP neural network of the existing prediction models.The soil moisture is predicted by simulating weather forecast with measured weather data,and the vertical average soil moisture measured at 40 cm in Shuangyang of Changchun City is used for verification and testing.The results show that the BAS-BP neural network has faster convergence speed and higher prediction accuracy than the traditional BP neural network.The average relative error of single-day forecast is 0.3,the average relative error of three-day forecast is 0.62,and the average relative error of five-day forecast is 0.92.At the same time,comparing with GA-BP and PSO-BP models,it is found that BAS-BP has better robustness.Combined with weather forecast data,it can accurately predict soil moisture changes in the next five days.Finally,a field moisture conservation strategy is proposed.This strategy is based on the aforementioned soil moisture prediction inverse model,combined with the weather data in the next five days to calculate the soil moisture threshold,and based on the obtained threshold to determine the irrigation method,and then use simulation to obtain a high-precision and high-precision The stable fuzzy PID algorithm combines the aforementioned work to establish a field water-saving irrigation system based on an embedded system.The system can predict soil moisture and modify the expected value of irrigation according to the predicted results and irrigation strategy settings,and use fuzzy PID control for precision Irrigation.Field tests and comparisons with traditional irrigation strategies show that water-saving strategies have good practical effects.While the water-saving effects are obvious,it is easier to maintain soil moisture in a suitable range.