Water stress detection using hyperspectral thermal infrared remote sensing

Auteurs

M. Gerhards, G. Rock, M. Schlerf, T. Udelhoven, and W. Werner

Référence

in 7th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, WHISPERS 2015, Tokyo, Japan, 2-5 June 2015, vol. 2015-June, art. no. 8075366, ISBN 978-146739015-6, 2017

Description

During summer 2014 a water treatment experiment was adopted on potatoes plants (Solanum tuberosum L. Cilena, n=60) with the objective to detect plant water stress. Therefore, three different sensors were applied, a hyperspectral and broadband thermal infrared (TIR) camera to measure canopy temperature, as well as a leaf porometer to measure stomatal conductance. The results of this study show that water stress can be detected from 2 days after stress based on stomatal conductance measurements using a descriptive t-test (p=0.044∗ at 5% level of significance). Applying the prominent crop water stress index (CWSI) based on canopy, dry and wet references temperatures, control and treatment can be significantly separated starting 8 days after stress (p<0.001∗∗∗) using both hyperspectral and broadband data. Thus, our hypothesis of an earlier and more accurate water stress detection using a hyperspectral TIR system must be falsified for this study.

Lien

doi:10.1109/WHISPERS.2015.8075366