Development Of Variable Rate Irrigation System Based On Uav Remote Sensing And Ground Monitoring Data

Variable irrigation is a modern irrigation technology aiming at "saving energy and increasing production".Variable rate irrigation system can cope with different irrigation demands in space changes,thus improving the utilization efficiency of irrigation water.The overall application level of variable irrigation technology in our country is relatively low,the system cost is high,and the difficulty of sensor layout has become the main factor restricting the development of variable irrigation technology.This research integrates advanced remote sensing system of unmanned aerial vehicle,fuzzy decision theory and control system,and develops a variable irrigation system.The main research contents and results of this paper are as follows:(1)According to the working principle of center pivot Sprinkler,the variable irrigation control system had been designed and implemented.The system consists of a remote monitoring terminal,a control system base station and an electromagnetic valve controller,wherein the main components of the remote monitoring terminal are a computer monitoring interface and a data transceiver,and the main control chip of the control system base station and the electromagnetic valve controller is STM32F103C8T6.In the system design,the base station of the control system can control the opening and closing time of the solenoid valve group according to the duty-cycle control map transmitted from the remote monitoring terminal,thus realizing variable irrigation.In addition,the system integrates a soil moisture sensor,and according to the change of soil moisture,the control system is closed when the soil moisture content at all monitoring points is greater than 40%.In the test of control system,the change trend of irrigation amount in different duty cycle control areas on site is basically consistent with the change of the set duty cycle.The irrigation amount when duty cycle is 1 is 583 g,Based on the average irrigation volume when the duty cycle is 1,the average absolute error of irrigation amount in the control system was 15.15%,and the average actual irrigation amount was 9.29% higher than the set irrigation volume.(2)The numerical value on the remote sensing image is selected by geographical position coordinates,which is not convenient for control system.Since the management area of center pivot Sprinkler is fan-shaped,the remote sensing data are stored and averaged in different areas according to the shape of the management area,which can simplify the design of the control system.This research combined GDAL and Numpy toolkits on python development platform to develop a remote sensing image analysis tool suitable for centerpivot Sprinkler.According to different resolution requirements and coordinates of central,the remote data of irrigation area is divided into fan-shaped management areas with different radii and equal central angles.In addition,the tool can set the "non-number" value of the image to zero,so that the "non-number" does not participate in the decision-making,and reduce the influence of erroneous values occurring in the splicing process on the decision-making results.(3)We designed a fuzzy system to make decisions on remote sensing data and finally generate a duty cycle control chart.In this study,two fuzzy decision systems were designed based on Matlab and Python,which can generate duty cycle control maps.The first fuzzy decision-making system is based on a multispectral inversion model.The Crop water stress index(CWSI)and Crop coefficient(Kc)are obtained by calculating the band data of the management area.Then the reference evapotranspiration(ET0)is calculated by referring to the FAO-56 evapotranspiration calculation formula,and the actual crop evapotranspiration(ETc)is further obtained.CWSI and water variation are used as fuzzy inputs,and Mamdani fuzzy reasoning is carried out on crop water demand according to fuzzy rules and fuzzy sets established by the empirical model.Finally,the duty cycle of the solenoid valve is obtained by means of central average deblurring.The second fuzzy decision system uses multispectral data to calculate vegetation index,uses vegetation index and original thermal infrared data as fuzzy inputs,determines fuzzy sets according to the maximum values of vegetation index and Digital number(DN)in the experimental area,and obtains duty cycle of solenoid valve by Mamdani inference method and means of central average deblurring.In addition,two irrigation time setting methods are also introduced for the system.In the decision-making system test,The results showed that when the water stress in the different treatment areas is more than 10% under the condition of the same planting structure and bare ground elevation,the output results of the two systems had good decision gradients.The decision accuracy of the fuzzy decision system based on original data was 90%.(4)Based on Py Qt platform,the variable irrigation decision system software was developed using python and matlab language,which could realize remote sensing data processing,visualization of solenoid valve duty cycle control chart and 3D analysis of local water shortage information.The software covered all contents of the decision system and all functions run well.