Autonomous Vision-based Unmanned Aerial Spray System with Variable Flow for Agricultural Application

The study focuses on implementing an independent aerial sprayer system with vision sensors for autonomous agrochemical applications using advances in precision agriculture. Poor planting management often leads to unpredictable vegetation vigor in croplands, resulting in wasted agrochemical inputs. Achieving precise spraying flow control using typical UAVs over crop fields is challenging. This study's aerial sprayer system can evaluate crop health using a visual-based RGBVI calculation, made possible by incorporating an RGB camera and a small computer. The system measures VIs, determines plant health, and controls the input flow rate. The autonomous vision-based aerial spraying system responds by detecting the real-time interpretation of the Red Green Blue Vegetation Index (RGBVI) for low and high vegetation vigor. It regulates the flow rate by generating a square pulse wave with varying duty cycles. The system uses an 850 mm hexacopter with a 3-kg liquid solution tank that can spray up to 800 ml/min, establishing a direct relationship between varying RGBVI and the duty cycle. When the duty cycle is between 50% and 100%, nozzle flow ranges from 2.85 mL to 5.78 mL per 30 seconds. A comparison was made on the planned flight routes and trajectories flown to assess the drone's autonomous flight capabilities in following the terrain. The drone sprayer could follow its intended path autonomously with an average deviation of 6.72cm and a terrain-following error of only 0.2%. © (2023), (International Association of Engineers). All Rights Reserved.