Distancebased Control System for Machine Visionbased Selective Spraying

For effective operation of a selective sprayer with realtime local weed sensing, herbicides must be deliveredaccurately to weed targets in the field. With a machine visionbased selective spraying system, acquiring sequential imagesand switching nozzles on and off at the correct locations are critical. An MS Windowsbased imaging system was interfacedwith a realtime embedded selective spray controller system to accomplish control tasks based on distance traveled. Amachine visionbased sensing system and selective herbicide control system was developed and installed on a sprayer. A finitestate machine (FSM) model was employed for controller design, and general design specifications were developed fordetermining the travel distance between states. The spatial application accuracy of the system was measured in the field usingartificial targets. The system operated with an overall hit accuracy of 91% with no statistical evidence of hit accuracy or meanpattern length being dependent on vehicle speed. Significant differences in pattern length variance and mean pattern widthwere detected across speed levels ranging from 3.2 to 14 km/h. Spray patterns tended to shift relative to the target at highertravel speeds.