This AI-enabled smart sprayer system for raised plasticulture beds and row-middle weed control delivers precise herbicide spray in real-time while maintaining accuracy under extreme weather conditions. Weeds and diseased vegetation create ongoing challenges in agricultural production. By competing with crops for sunlight, nutrients, and water, ineffective weed control can reduce overall yield and crop quality. As growers increasingly adopt precision agriculture technologies to address these issues, the global smart crop scouting and smart spraying market, which is valued at $5.33 billion in 2024, is projected to reach $19.55 billion by 2034, reflecting a compound annual growth rate of 13.86% over that period . However, existing sprayer systems are optimized for flat, soil-based rows and struggle to operate in the uneven, sandy terrain of raised plasticulture. Therefore, there is a clear need for precision agriculture technologies that can reliably target weeds on raised beds and row middles while withstanding extreme weather conditions.
Researchers at the University of Florida have developed an AI-enabled smart sprayer system for raised plasticulture beds and row-middle weed control, maintaining precise, real-time herbicide applications even under the extreme heat, humidity, and wind typical of Southwest Florida. This invention integrates machine vision, machine learning, pump pressure regulation, and robotic mobility into a unified platform tailored for specialty crop production. By leveraging a detection model, the system activates nozzles only at verified weed locations, eliminating drift and overspray on uneven, sandy terrain. This directly addresses the pressure-control and terrain-adaptability gaps of existing see-and-spray solutions, opening new opportunities in precision-ag applications.
This AI-enabled smart system supports precision weed control and selective agrochemical or foliar nutrient delivery in specialty crops cultivated on raised plasticulture beds and row middles
The platform is designed to travel along raised plasticulture beds and row middles using terrain-capable tires and a structural framework configured for these environments. The system incorporates image capture devices that stream real-time imagery to an onboard computing system containing image processing, a trained object detection model, and a spray control module. The detection model, which was trained on weed types common to targeted crop systems, including broadleaf, sedge, and grass, identifies and localizes agricultural subjects and generates site-specific spray commands. These commands activate electronically controlled nozzles supplied by multiple DC pumps and proportional control valves. Integrated proportional-integral controllers regulate pressure and flow, maintaining consistent spray output while reducing drift and off-target application. The platform operates using a battery-powered configuration (48V 100A described) and includes protective structural elements that shield electrical and hydraulic components from wind, rain, heat, and moisture during field operations.