Research On Variable Spraying Weeding System Based On Machine Vision Technique And Fuzzy-PID Compound Control

Farmland weeds are the enemy of agricultural production,seriously affecting the growth space of crops the growth space of crops.Cleaning up weeds in the field timely and effectively is of great significance to improve the yield and quality of crops.Although chemical herbicides can effectively control weed hazards,extensive chemical spraying not only wastes pesticides,but also causes serious environmental and ecological pollution.In order to improve the utilization rate of pesticides,reduce environmental pollution and realize the function of variable spraying of weeds,After summarizing the domestic and foreign research on variable spraying technology and its latest development,this paper,taking weed cleaning in maize field as the research object,puts forward a set of variable spraying herbicide solutions.The main research contents and results are as follows:(1)A variable spraying weeding system was designed based on machine vision and fuzzy-PID composite control technology.The system mainly includes three parts: weed information acquisition module,variable spraying controller and software monitoring system.The weed information acquisition module is composed of GPS differential positioning system,speed sensor,single chip computer,camera and PC computer.It is used to collect the density and position information of weeds in the field.The variable spray controller is composed of microprocessor,flow sensor,pressure gauge and solenoid valve,which is used to realize real-time variable control function of spray quantity.The software monitoring system is used for human-computer interaction function,which can realize system parameter setting and system working status monitoring.(2)The Nikon D810 monocular digital camera was used to realize field image acquisition,and its installation position was designed.And several processing algorithms involved in the process of image processing were studied and analyzed.The appropriate graying algorithm,filtering algorithm and threshold segmentation algorithm were determined by the comparative experiment with the collected field images,which optimized the system structure and improved the system performance.(3)A Fuzzy-PID compound control algorithm was designed.The algorithm combines the advantages of fuzzy control and PID control.It compensates for the deficiency of PID control with the good non-linearity and good dynamics of fuzzy control.At the same time,by using the good steady-state characteristic of the PID control,it solves the problem that there exists a control blind zone near the steady-state equilibrium point of the fuzzy control.(4)Field test results showed that the relative control error of the system is less than 4.5% at the operating speed of 2km/h,which meets the design requirements.