Study On Soil Moisture Fast Measuring Technology And Soil Moisture Forecast Model

The soil moisture content of soil moisture, restricts the nutrients in the soil and the dissolution of minerals and transfer, has a vital role on the growth of crops. Only timely and accurate grasp of the soil moisture loss to timely drainage irrigation condition can ensure the normal growth of crops.This paper is mainly study of the rapid measurement technology of soil moisture and the soil moisture forecast model.First, this article based on the frequency domain reflection-FDR type soil moisture sensors, designed a temperature compensation of the portable soil moisture meter, which in addition to the traditional apparatus can only determine the function of soil moisture, its built-in GPS module can positioning measurement location latitude and longitude information at any time; Chosen CAT24WC256 as storage chip, which can maximum storage 2400 groups of 50 bytes of data; The built-in GPRS/GSM module, can transmission the measured data to the backend database and mobile phone users timely with wireless; The instrument also has multifunctional options menu, including real-time monitoring, storage process, monitor setting, number setting and so on. This meter is more appropriate and in accordance with the actual needs of agricultural production, has been in preliminary application.Secondly, through comparative tests, system and comprehensively study of the soil texture and temperature on the FDR type soil moisture sensor measurement accuracy, it is concluded that FDR type soil moisture sensor test has a bigger difference in different soil texture, and has temperature effect obviously; Preliminary concluded three correction function of soil texture and temperature compensation function, This calibration method is simple, practical and feasible, and the introduction of which greatly improved the accuracy of the instrument and the applicable scope.Finally, chose the drought-prone region FeiDongXian which is the jianghuai watershed as the study area, use the lots of data by soil moisture monitoring and small stations, establish the soil moisture forecast model based on BP neural network. Choose soil moisture, temperature, humidity, radiation, rainfall as the input and output of the model, two kinds of forecast model is set up respectively, forecast the soil moisture on the second day and five days after, discusses the BP neural network in the application of the short-term soil moisture forecast. The experimental results show that the second day after the forecast accuracy is higher than 5 forecast accuracy, the main reason is that the efficiency of the soil moisture and is strongly influenced by artificial factors and rainfall, so in the absence of irrigation and rainfall drought season predicted by this method can obtain better effect and in order to get high precision of prediction result, a large number of continuous accurate data is indispensable. Overall, the BP neural network in short-term soil moisture forecast obtained higher forecast accuracy, illustrates the feasibility of this method to be applied in prediction of soil moisture, providing decision basis for guiding agricultural production.