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

Implementing an aerial sprayer system with vision sensors that allow farmers to apply agrochemicals remotely is one of the most significant transformations in precise agriculture. However, cropland usually has random vegetation vigor caused during the planting management stage, resulting in agrochemical input waste. When spraying over crop fields using a typical UAV, it is not easy to control the accuracy of the spraying flow. This research develops a vision-based aerial spraying system that measures the Green Leaf Index (GLI) and identifies low and high vegetation vigor while spraying precise agrochemical applications. A small single-board computer serves as the core processor of the control system. With vision sensor technology, the system converts Green Leaf Index (GLI) to get quantitative value through real-time interpretation. It generates a square pulse wave with varied duty cycles to control the flow rate. An 850 mm hexacopter with a 3-kg liquid solution tank could spray 800 ml/min. The spray volume difference between the two nozzles is modest at 1.73 percent, and a minimum distance (D) of 0.9 meters is required to prevent gaps and overlaps. GLI values for the outdoor tests of the vision-based aerial system reveal a varying 0.0 to 0.4 vegetation index. The design provided a direct relationship with varying GLI to the duty cycle. When the duty cycle is between 50 and 100 percent, the nozzles’ flow range fluctuates from 2.85 mL to 5.78 mL per 30 seconds. Implementing GLI’s onboard vision-based calculation allows for the variable flow and eventually improves yield while reducing agrochemical input.