Abstract:
Our FYP project proposes the development of a variable rate sprayer system that can be mounted on an agriculture drone. The objective of this project is to design and implement a system that can accurately and efficiently apply crop protection products in a precise and targeted manner. The variable rate sprayer incorporates an electronic flow control system, which will be responsible for adjusting the flow rate of the sprayer in real-time based on the crop density data.
The crop density data for a sample field is calculated using Deep learning classifier, ResNet 50 which classifies the crop field based on the pest percentage. The classifier is successfully trained up to an accuracy of 82% to classify The field into four categories, with pest below 25%, 25-50% pest, 50 – 75 % pest and above 75% pest.
The Pixhawk flight controller is capable of controlling the flight dynamics of the UAV and the flow rate of the sprayer during the same flight. Therefore, the need for a secondary controller for sprayer is eliminated.
The CFD simulations of the variable rate sprayer at the nozzle are performed to analyze the flow physics inside the nozzle. The results are validated by the experimental validations and the variable flow rates are found to be in line with CFD results.
Overall, the project, when delivered on a large scale is able to reduce the labor time to manually the agriculture field, by about 60-70% and subsequently the costs of spraying is reduced
