In order to overcome the shortcomings of unreasonable mixing of traditional pesticide spraying methods and the complexity and high cost of spraying control, a variable spraying system with independent supply control of water and pesticide was designed in this paper. Firstly, the SK mixer with 3 units was determined by analyzing the performance of SK, SX and SH mixers. Then, the test schemes such as nozzle speed and nozzle height were designed. When the nozzle speed was 11 750r/min (Duty cycle was 75%), the amplitude angle was 112.6, and the nozzle height was 1.0 m, the number of particles per unit area of crops reached more than 48 pieces/cm2, and the error of pesticide concentration was not more than 3%. Finally, taking the wheat sheath blight as the control object, the spraying experiments were carried out on the manual operation area, the UAV operation area, the variable spraying area and the blank area. The results showed that the sum of grade 1 and grade 2 growth conditions in the variable spraying area accounted for 97%, and the control effect of wheat sheath blight at returning green stage was ideal, which was higher than that in other areas after 15 days and 30 days of spraying. 

The development of traditional agriculture is restricted by population aging and high labor intensity. With the development of information technology and intelligent control technology, the emergence of unmanned agricultural vehicles may fundamentally solve this problem. Based on the introduction of the current development of unmanned agricultural vehicles, this paper focuses on the development status of environment perception technology of unmanned agricultural vehicles and compares and summarizes different technologies.It is pointed out that the development and application of environment perception technology of unmanned agricultural vehicles cannot follow the development direction of fine agriculture, and the research on the structure of the agricultural production environment is lacking. In order to reduce the cost and popularization, the paper points out the development trend of the environment perception technology layer of unmanned agricultural vehicles, such as the scalability.