Project

Drone-based drought stress detection

Potato crops are increasingly facing drought stress. One fast way to detect drought stress is to measure the canopy temperature, by capturing thermal images with drones. However, these measurements are strongly influenced by the environment and additional work is needed to ensure accurate temperature measurements.

Project description

Due to the effects of climate change, Dutch potato crops will experience increasing drought stress in the future. Potato plants experiencing drought tend to close their stomata to reduce water loss through transpiration. At the same time, this restricts the intake of CO2 and, consequently, photosynthesis. However, not all potato varieties are equally affected. Some varieties possess traits which make it more resilient against drought stress, e.g., by developing a deeper root system, enabling them to keep their stomata open and maintain photosynthesis.

A side effect of transpiration is evaporative cooling. In theory, this leads to cooler canopy temperatures for crops with sufficient water availability compare to those under drought stress. The relationship between water availability and canopy temperature has been used to define the crop water stress index, which uses the canopy temperature as a proxy measurement for the limitation on photosynthesis and crop productivity.

Temperature is relatively simple to measure using
thermal infrared cameras. Mounting these cameras on drones allows for
high-throughput phenotyping of canopy temperature, and with it drought stress. To
serve as an effective indicator the temperature measurement has be accurate, however,
thermal infrared sensors are easily influenced by the environment, and even a
passing cloud can have a strong effect on the measured value. Therefore, before
the data can be used for analysis, it needs to be calibrated to minimize noise.

Objective and methods

Objectives

The project will compare different methods of calibrating drone-based thermal infrared images, e.g., using reference surfaces or machine learning.

Methods

Thermal image data has been collected on a large-scale potato field experiment with different irrigation schedules and varieties. There are additional measurements, e.g., on the temperature of reference surfaces and canopy temperature using ground-based sensors. There is the option of collecting additional data in 2025 by flying the drones yourself (we can help you getting the necessary drone license and training).