Drone-mounted optical sensing system for real-time targeted agricultural input application

Overview

This project develops a drone-mounted precision agriculture platform that combines optical sensing, onboard computing, and synchronized spraying to apply agricultural products such as fertilizer only where and in the amounts needed. Optical sensors measure indicators like soil nitrogen content in real time as the drone passes over the field. An onboard computer processes those readings, calculates the appropriate application rate, and triggers a sprayer that is precisely synchronized with the drone's travel speed. Because drones move faster than ground-based application vehicles, the system promises faster field coverage and more efficient input use, reducing waste and environmental impact.

Technical specifications

• Optical sensing: An optical sensor measures voltage readings that are converted into the Normalized Difference Vegetation Index (NDVI), a standard indicator of crop health and nutrient status such as nitrogen content.
• Onboard processing: Python code processes the sensor readings into fertilizer application rates on an embedded computer such as a Raspberry Pi, keeping computation time low enough for real-time decision making.
• Synchronized spraying: Knowing the front-to-back distance between the sensor and the sprayer and the drone's horizontal speed, the system calculates the precise trigger time. A wait time is applied so that the sprayer activates exactly when the drone is over the targeted area.
• Variable-rate application: The sprayer pump is pulse width modulated (PWM) based on the computed application rate and drone speed, achieving the calculated target dose on the fly.
• Flight control: The system uses open-source ArduCopter firmware and the MAVLink protocol, with DroneKit-Python development tools, keeping the software stack free and customizable.
• Hardware components: A custom-built quadcopter drone, optical sensor, sprayer kit with tank and PWM pump and nozzles, and flight controller, with on-campus lab access for assembly.

Technology readiness level

The project is at an early development stage, moving from concept to bench and field validation. Planned validation includes on-the-bench testing of the computation pipeline on the Raspberry Pi to confirm that processing time is short enough for real-time operation, followed by field testing with a custom-built quadcopter to verify sensor-to-sprayer synchronization and variable-rate application under real flight conditions. Funding is being sought to support a student researcher and to acquire the drone components, optical sensor, and sprayer kit needed to complete assembly and testing.