Design And Implementation Of Target Spray System Based On Displacement Detection And Crop Identification

Target spray technology is an important part of intelligent plant protection.In order to improve the utilization rate of pesticides,it is necessary to adjust the spraying scheme adaptively according to the speed of the crop and the speed of the sprayer,so as to reduce the environmental deterioration caused by excessive spraying.With the progress of MEMS semiconductor processing technology,combined with the traditional integrated circuit technology,MEMS inertial sensors are small and light in weight,so they are widely used in aviation,aerospace,robotics,consumer electronics and other fields.According to the acceleration principle and coordinate transformation method,a high-precision three axis MEMS acceleration sensor ADXL345 is adopted,and I2C bus communication mode is used to design the displacement detection module.The acceleration is converted to displacement using the integral calculation principle.In order to reduce the measurement error of the sensor,the error source of the ADXL345 acceleration sensor is analyzed,and the mathematical modeling of the error is carried out.In the displace.?ment detection module,a filter link is added before double integration.Besides the static error compensation,the adaptive EMD filtering method is also selected for the random error of the sensor.On the basis of intelligent target spray system,combined with MEMS acceleration sensor module,in the Microsoft Visual Studio 2010 environment,the algorithm of target spray displacement detection system is implemented,and the error of sensor is analyzed.In the target spray equipment,the detection of crops is completed by using the common visual library function of OpenCV.in order to get the green target in the image,we first choose the super green method to process the grayscale.Then we use the maximum variance method between the classes to segment the threshold automatically,which can effectively separate the green crops from the background.The host computer makes the spray scheme according to the information obtained by the machine vision,and controls the lower computer by sending the signal through the serial port.After the hardware and software design of the main components of the system is completed,the alarm module is added to make the system more intelligent.Finally,the test platform of the system is built.Field simulation experiments show that the EMD filtering algorithm effectively suppresses the random error,improves the accuracy of displacement detection,and meets the demand of speed adaptive adjustment for spraying timing.The intelligent spray system has reached the level of practical application.