Research On Visual Inspection Of Operation Quality And Variable Spraying Control Of Sprayer

In order to improve the technical level of plant protection machinery and equipment,machine vision technology and image processing method are used to carry out visual inspection on the working quality of the sprayer.Based on the visual inspection method,the variable spraying control system of the sprayer was designed.The method can effectively identify the leaf adhesion droplets and improve the accuracy of the detection of the deposition distribution of droplets on the leaf surface.Variable spraying control system can achieve precise application,reduce pesticide use,and improve the efficiency of pesticide use.In this paper,a foliar fog droplet acquisition device with smart phone as the core was constructed to collect fog droplet images for plant protection operations in the field.Nine groups of spray detection leaf samples were selected,and the features of leaf region and fog droplet region were extracted by image processing algorithm based on gray morphology.A droplet type recognition method based on contour firmness is proposed and compared with the traditional droplet recognition method based on roundness and shape factor.Based on the results of cyclic open operation of attached droplet,the traditional distance transform watershed image segmentation algorithm is improved to detect the distribution of foliar droplet deposition.The control scheme and program of variable flow rate and height spray are designed.According to the test results,the spray flow is adjusted to achieve variable flow control.The optimal spraying height range is set in advance,so that the spraying rod of the spraying machine can automatically adjust the spraying height in the best range,and the variable height spraying control is realized.Software simulation and simulation control are carried out respectively.The results show that the water drop coverage of the improved watershed detection algorithm is 31.93% and 33.43% higher than that of the OTSU algorithm and the block threshold algorithm,respectively.The average correct rate is 97.23%.The accuracy of the solid based droplet type recognition method is 21% and 5% higher than the roundness and shape factor based on methods,respectively.The average relative error between the deposition density and manual statistics is 6.61%,and the average correct rate is 94.74%.In the simulation and control of variable flow spray,the motor can be directly rotated to 0° or 90° according to the opening adjustment requirements,and the rotation angle of 10° can be adjusted each time to achieve the purpose of regulating the valve opening.In the simulation and simulation control of variable height spraying,the ultrasonic sensor can control the rotation direction of the motor with 0.2cm accuracy according to the preset height range of 30 ～ 50 cm,in order to maintain the nozzle in the best working height range.The identification method of attached fog droplets based on contour fixation and the improved image segmentation watershed algorithm proposed in this paper can provide technical support for field operation quality image detection of plant protection equipment.The research of variable spray control can provide reference for variable spray control of spray machine.